Archives | VDGOOD

Volume: 1- Issue: 2

a multi-level security system for larceny intimation using internet of things(iot)

¹Ms.Priyanka S , ²Roshan Sabeqah , ³A.Siva Ranjani , ⁴B. Monica Jenefer

¹Student, Computer Science and Engineering, Meenakshi Sundararajan Engineering College, India.
²Student, Computer Science and Engineering, Meenakshi Sundararajan Engineering College, India.
³Student, Computer Science and Engineering, Meenakshi Sundararajan Engineering College, India.
⁴Associate Professor, Computer Science and Engineering, Meenakshi Sundararajan Engineering College, India.

Abstarct [+]

First and foremost, iot are Smart, connected appliances. An intelligent house is programmed to save energy and make your life a more convenient one: Alarm clocks will be synced with traffic apps , heating systems will be synced with external temperature sensors, which will be synced with cost evaluations; lighting will react as we enter a room, as might our coffee makers. There are plenty such homes that already exist. There are various fields were iot has placed its strong firm in the phase of time. This paper deals with the technological advancements in the phase of (iot) internet of things in the field of security. The demand for iot has become more and more reliable in the development of security. As security plays a vital role in abolishing data breach, its very much mandatory to safe guard once’s information .Therefore this project helps in the multilevel security enhancement which accounts 3-tier authentication of information that includes biometrics, number lock, hand held devices.

| Download

data analytics for crop harvesting using logistic regression algorithm

¹Dr.Pritto Paul P , ²Sakthi Ganesh M , ³Vamsi Krishna C , ⁴Girish Kumar R

¹Assistant Professor, Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

In recent times, people step back from harvesting crops due to various factors like yield obtained in specific season over a region and return of investment. In addition to this, modern youth are unaware of ideologies implemented in agriculture. In order to analyze the data sets of agriculture, relational database(SQL) is being used currently which has many limitations. The relational database consumes more time for processing huge data sets. Maintenance cost is also very high. To overcome this problem, Hadoop is used for analysing huge data sets since it enables businesses to run applications on thousands of nodes involving terabytes of data supporting parallel execution. Hadoop framework primarily consists of two important tasks namely Hadoop Distributed File System(HDFS) and Map Reduce. For validating the data sets analysed by Hadoop, a binary classification algorithm called Logistic Regression is used which is based on the concept of supervised learning. Logistic Regression measures the relationship between the dependent variable and one or more independent variables by estimating probabilities using it’s underlying logistic function. The hypothesis tends to limit the logistic function between 0 and 1. Logistic Regression algorithm thus uses a predictive analysis technique based on probability.

| Download

defending of intra-domain ip spoofing in software defined networking

¹Mr.Ajay V , ²Arun K S T , ³Sudarshan S , ⁴Vijayalakshmi M

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Assistant Professor, Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

The increasing security threats and the lack of secure authentication and encryption leads to illegal access in the network traffic. The Internet Protocol (IP) was the source of Internet transmission but an inadequate authentication technique paves a way for IP spoofing. Since, IP spoofing initiates more attacks in the network, a secure authentication technique was required to defend against spoofing attacks. The prevention of attacks such as DDOS and man in the middle attack are so hectic and the primary cause of them is IP spoofing, as it is important to prevent the IP spoofing. Existing system is inefficient because of huge overhead involved in implementing the export policies to be followed by each node or AS to defend the IP spoofing. In order to provide the solution, the Neighbour Authorization (NA) Algorithm is used to defend IP spoofing attacks with the help of SDN (Software Defined Networking) by constructing the NA table. This technique authenticates IP address of each host using NA algorithm. The algorithm works by validating NA table constructed by each node. The NA table is transmitted securely as it plays an important role in authentication process. The files send through the network are encrypted using AES (Advanced Encryption Standard) for security reasons so that the files cannot be seen by others, except the person with the access to the file.

| Download

detecting fetal brain tumor using convolutional neural network

¹Mrs.Shanmugha Priya M , ²Aravindhini N D , ³Bhagavath Meena S S , ⁴Vaishnavi V

Abstarct [+]

Identifying fetal brain tumor at early Gestational Age(GA) of the fetus is significant as 3 out of 1000 women are pregnant with abnormal fetal brains. One of the abnormalities found in the fetal brain is a tumor. Detecting fetal brain tumor at the early stage with machine learning techniques can improve the quality of treatment and precautions. Early detection of the fetal brain tumor will indicate how the pregnancy can be managed, possible treatments that can be taken, and help parents understand and be prepared for dealing with the tumor. Fewer studies considered the use of machine learning methods to identify defects existing in the fetal brains. They constructed a classifier for predicting the SGA abnormality after the birth of the fetus (becoming a neonate). But, this method has several limitations. The proposed algorithm is capable of detecting and classifying a tumor from MRI images with a wide range of fetal gestational age (GA) using a flexible and simple method with low computational cost. The novel proposed method consists of four phases; segmentation for brightness correction, enhancement using dual-tree complex wavelet transform, feature extraction using Gray Level Co- occurrence Matrix and classification using convolutional neural networks.

| Download

gesture-based digital art - drawing in the air using object tracking algorithm

¹Ms.Goga pooja pravallika , ²Pujitha P , ³Ramya R , ⁴Bhagyalakshmi M

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

Digital revolution is one of the major trends in the industry. Crafting is one of the sturdy jobs then but now it becomes easy. Here we introduce digital art drawing in the air through hand gestures and using a camera capturing the digital art. We are giving different hand gestures to draw art. Here the gesture captured by a camera is displayed as digital art. Gesture recognition is a technology that is used to identify human gestures with the help of object tracking algorithms and centroid estimation. Gesture identification determines the hand, tracks the object movements and thus brings the information about object position and flux of the fingers. Thus these application allows the user to draw on any kind of surface by tracking the fingertip movements of the user’s. The images that are drawn by the user can be stored on any other surface. The user can also muddle through various images and drawing by using hand gesture movements. Shading markers are set at the tip of the fingers of customers. Here we create digital art entirely by executing algorithms that are coded into computer programs and can be regarded as the computer's native art form. Digital art can be made by tracking the colors on fingertips and digital art will appear on screen according to our hand movement.

| Download

extraction of secured e-health record using hyperledger fabric

¹Mr. Pritto Paul P , ²Pavithra I , ³Gayathri J

Abstarct [+]

In the digital healthcare era, it is of the utmost importance to harness medical information scattered across healthcare institutions to support permissioned Blockchain and achieve personalised healthcare. However, cyber infrastructure boundaries of healthcare organisations and privacy leakage threats place obstacles on the sharing of medical records. Blockchain, as a public ledgers characterised by its transparency, tamper-evidence, data privacy, and decentralisation, can help build a secure medical data exchange network. Existing work mainly pay attention to Ethereum blockchain and the decentralised interplanetary files system (IPFS). It has not been so successful due to security and privacy flaws observed over time. In this paper, we propose a framework by using effective public key encryption mechanism and Hyperledger fabric to ensure the security of the data. The use of permissioned blockchain-based approach is suggested, which ensures that the blockchain is owned by a governing body and no unauthorised access can be made from outside. We also propose an authentication schemes for the blockchain-based EHRs. This paper claims to apprehend the security and data management challenges in blockchain and provides improved manifestation of medical data sharing.

| Download

detection of oil spilled regions in ocean from synthetic aperture radar images using fuzzy c-means algorithm

¹Mr.Devanathan R , ²Viyagappan A , ³Vishal rajan M , ⁴Mrs. Rajalakshmi S

Abstarct [+]

The Oil spill is one of the main environmental hazards in ocean pollution. At present, the problem of marine pollution caused by oil spill accidents is increasingly serious. Rapid and accurate automatic recognition of SAR images provides an important prerequisite for the handling and decision of oil spill accidents. In our existing system,change detection algorithms has been used, which shows oil spilled area by processing sequence of images.Detection of the oil spill in ocean is identified by comparing images acquired at different times. The specific algorithms used are the correlation coefficient change statistic and the intensity ratio change statistic algorithms. However,algorithms used in the existing system consumes more processing time and is more sensitive to the changes in the areas which are represented by bright pixels compared to the changes in areas represented by dark pixels. Therefore, especially when the oil spill is close to land in an image, and also when contrast of the image is very low, these algorithms behave poorly. In our proposed system, the detection of oil spills in ocean is evaluated and improved with FuzzyC-Means algorithm.Initially,we have performed preprocessing such as grayconversion, filtering and image contrast enhancement. Afterwards, Fuzzy C-Means algorithm is used for image segmentation which is used to segment the oil spilled regions from other regions. Using this technique, image is segmented into number of clusters,based on feature extraction. It’s processing time and accuracy is better than existing algorithms,as it works more efficiently even when the contrast of image is very low. Then, the area of oil spilled region is calculated based on the pixels in binary image. By using Support Vector Machine algorithm, Classification is performed by taking area of oil spilled region as input training dataset,and it is used to classify whether the calculated oil spilled area is large or small.

| Download

detection of diabetic retinopathy by fundus image using convolutional neural network

¹Mr.Ajay Saravanan R , ²Ganesh Krishnan R , ³Visvarasu A , ⁴Ms. Revathi S

Abstarct [+]

The condition of the vascular network of human eye is a crucial diagnostic think about retinopathy. The fundus image segmentation is critical thanks to variable size of vascular vessels and potential presence of pathologies like micro aneurysms and haemorrhages. The Project proposes the Retinal image analysis through efficient detection of vessels and exudates for retinal vasculature disorder analysis. this will be helpful in detection of some diseases in early stages, like diabetes, which may be performed by comparison of the states of retinal blood vessels. Intrinsic characteristics of retinal images make the vessel detection process difficult. Here, we proposed a replacement algorithm to detect the retinal blood vessels effectively. The green channel are going to be selected for image analysis to extract vessels accurately. The Daubachies wavelet transform is employed to reinforce the image contrast for effective vessels detection. The directionality feature of the multistructure elements method makes it an efficient tool in edge detection. Hence, morphology operators using multistructure elements are applied to the improved image so as to seek out the retinal image ridges. Morphological operators by reconstruction eliminate the ridges not belonging to the vessel tree while trying to preserve the skinny vessels unchanged. so as to extend the efficiency of the morphological operators by reconstruction, they were applied using multistructure elements. an easy thresholding method along opening and shutting indicates the remained ridges belonging to vessels. Experimental result proves that the blood vessels and exudates are often effectively detected by applying this method on the retinal images.

| Download

detecting the negative comments in social forum using support vector machine

¹Ms.A. Saranya , ²R.Shreya , ³Dr.P.Visu

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Professor , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

In recent times, Social media has became very talked-about way for internet users to speak and interact online. Users spend many time on famous social networks (eg Facebook, Twitter, Amazon, Flipkart) ordering the commodity and reviewing. Unfortunately, this popularity also attracts a big amount of spammers who continuously expose to malicious behaviour resulting in inconvenience to the dealer who sell the commodity. Thus crawling the negative review from all the social forum for a specific commodity (search keyword) help to crawl all the user and their comments from the web site. The negative comments is chosen from the all the crawled site by regular expression matching of the negative terms that's enclose the backend. The crawled data is validated , classified using Support Vector Machine algorithm and Convolutional Neural Network to validate the input parameters and procure the result supported the regularization of hierarchical patterns and also the fake user is identified when one specific user is commenting a negative review for the identical dealer and within the same social site over the actual threshold value , then the user have to be reported to the actual site and blocked the URL allowing unaccess for the user to access the location.

| Download

detection and classification of plant leaf diseases using convolution neural networks

¹Mr.Nikil C A , ²Santhoh P , ³Roshan Kumar S , ⁴Mrs. Rajalakshmi S

Abstarct [+]

Crops suffering from various diseases can be a big turndown for crop yield. This can affect effective crop production, if left unnoticed. Hence, it is of great importance to diagnose the plant diseases at early stages so that appropriate and timely action can be taken by the farmers to avoid further losses. This project focuses on the approach which is based on image processing for detection of diseases of plants. In this paper, we propose a system which uses convolutional neural networks that helps farmers to identify any possible plant disease by uploading a leaf image to the system. The system consists of a set of algorithms which can identify the type of disease. Input image given by the user undergoes several processing steps to detect the disease and results are returned back to the user on a user interface, for Example: Computer screen.

| Download

detection of bike riders with no helmet using haar cascade algorithm

¹Mr.Ganesh T V , ²Yogesh S , ³Pravin K V , ⁴Mr. Ganesan R

Abstarct [+]

Bike accident have been quickly developing during the time in numerous nations. The helmet is the safety equipment for motorcyclists however numerous drivers don't utilize it. We developed an helmet detection method combining classification and cluster. Helmet detection is an important, yet challenging vision task. It is a critical part in many applications such as traffic surveillance. Our proposed method work is as follows, Pre- processing, Feature Extraction and classification. We demonstrate our proposed work by using surveillance traffic videos. Finally , our method will classify whether the person is wearing helmet or not.Haar Cascade is a machine learning objectdetection method used to identify by image or object. It is a machine learning approach where the set is based on the lot of positive and negative images then can be used to detect specific objects. It helps to find whether the person is wearing helmet.Feature descriptors used to cryptograph into a series of numbers and act as a finger print to a series of number that can be used to differentiate one feature from another. Ideally this information isconstantunder image transformation, so we can able to identify it easily even if the image is transformed in some way. It helps to recognize the vehicle number from the number plate of the victim.As far as the robustness and effectiveness are concerned, our method is better than the existing algorithms.It automatically recognize the person not wearing the helmet and capture the vehicle number from the number plate and fine amount will be sent to the particular mail id with picture proof.

| Download

crime incidents detection using support vector algorithm

¹Mr.Ravikumar B , ²Mohamed Fahad Ali Abbas J , ³Githendra Vishal B , ⁴Vasudevan T

Abstarct [+]

The objective of this project is to tackle a vital issue in the society - Crimes. Analyzing and examining of crimes happening in the world will give us a Broadview in understanding the crime regions and can be used to take necessary precautions to mitigate the crime rates. Identifying Crime patterns will allow us to tackle problems with unique approaches in specific crime category regions and improve more security measures in society. Current studies show the reason of increase in crime rates is more in areas that are economically backward. In property crime will be a targeted in the upcoming decades. The following approach involves predicting crimes classifying, pattern detection and visualization with effective tools and technologies.Past crime data trends helps us to correlate factors which might help understanding the future scope of crimes.

| Download

detecting phising attacks using svm and random forest techniques

¹Mr.Muthuraja Sethupathy A , ²Nagaraj M , ³ , ⁴Mrs. Sridevi S

Abstarct [+]

Phishing is a form of fraud in which the attacker tries to learn sensitive information such as credentials of login or account information by sending as a reputable entity or person in email or other communication channels. The message contains malicious software targeting the user’s computer or has links to direct victims to malicious websites in order to trick them into divulging personal and information about financial, such as passwords, account IDs or credit card details. Phishing is known attackers, since it's easier to trick someone into making a short malicious link which seems justify than trying to stop the continuous process through a computer’s defense systems. The malicious links within the body of the message are designed to make it appear that they go to the spoofed organization using that organization’s logos and other legitimate contents. In the existing system a poorly structured NN model is used and unnecessary features are included for the classification process which may cause the system to be over fitted or under fitted and the usage of Naïve Bayers classification also leads to the poor performance of the system. We propose a learning based approach in which the system uses a optimal feature extraction which addresses the fitting problems and the usage of SVM and Random Forest algorithm also increases the accuracy of the system. We classify the Websites into 3 different types of classes: Benign, Spam and Malicious. By employing learning algorithms, our scheme achieves more performance on generality and coverage , when compared with blacklisting service. We classify the URL into, Benign which is a safe websites with normal services, Spam performs the act of attempting to flood the user with advertising or sites such as fake surveys and online dating etc and Malware which are created by attackers to disrupt computer operation, gather sensitive information, or gain access to private computer systems. Based on the above given information the websites which are harmful is blocked at the user computer. So the next time the user can’t able to access those websites

| Download

detection of hate speech and offensive language in twitter using sentiment analysis

¹Mr.Harish Krishna K , ²Thirunayan Manikantan R , ³Aravind Kumar D , ⁴Amirthavalli R

Abstarct [+]

The exponential growth of social media such as Twitter and other community forums has revolutionized communication and content publishing, but is also increasingly exploited for the propagation of hate speech and the organization of hate-based activities. The anonymity and mobility standard afforded by such media has made the breeding and spread of hate speech – eventually leading to hate crime – effortless in a virtual world that are beyond the realms of law enforcement. Existing methods in the detection of hate speech primarily cast the problem as a supervised document classification task. These can be divided into two categories: one relies on manual feature engineering that are then consumed by algorithms such as SVM, Naive Bayes, and Logistic Regression (classic methods); the other represents the more recent deep learning paradigm that employs neural networks to automatically learn multilayers of abstract features from raw data (deep learning methods). In this method we show that it is a much more challenging task, as our analysis of the language in the typical datasets shows that hate speech lacks unique, discriminative effects and therefore is found in the ‘long tail’ in a dataset that is difficult to discover. We then propose Deep Neural Network (DNN) structure as a unique feature extractors that are particularly effective for capturing the semantics of hate speech. Our methods are tested on the largest collection of hate speech datasets based on Twitter, and are shown to be able to outperform state of the art by up to 6 percentage points in macro-average F1, or 9 percentage points in the more challenging case of identifying hateful content. As a proxy to quantify and compare the linguistic characteristics of hate and non-hate Tweets, we also propose to study the ‘uniqueness’ of the vocabulary for each class.

| Download

emotion recognition using support vector machine

¹Mr.Sundar K , ²Sadagopan E.N , ³Chandran M , ⁴Aswin Raja S

Abstarct [+]

Human emotions are mental states of feelings that arise spontaneously rather than through conscious effort and are accompanied by physiological changes in voice muscles which implies recognises on voice. Some of critical emotions are happy, sad, anger, disgust, fear, surprise etc.Voice Recognises play a key role in non-verbal communication which appears due to internal feelings of a person that reflects on the voices. In order to computer modelling of human's emotion, a plenty of research has been accomplished. But still it is far behind from human vision system.In this paper, we are providing better approach to predict human emotions using deep Convolution Neural Network (CNN) ,Support Vector Machine (SVM) and how emotion intensity changes on a voice from low level to high level of emotion. This algorithm can be used in digital assistants like Amazon Alexa ,Google Home to recognise the user emotion and provide services based on their emotional state. The assessments through the proposed experiment confers quite good results and obtain accuracy may give encouragement to the researchers for future model of computer based emotion recognition system.

| Download

classification of brain tumor images using convolutional neural network

¹Mr.Navaneethakrishnan P , ²Anbarasan K , ³Ishindiran S , ⁴Krishna Parthasarathy

Abstarct [+]

Brain tumor can be defined as unnatural and uncontrolled growth in brain cells. Brain tumor classification is a crucial task to evaluate the tumors and make a treatment decision according to their classes. There are three categories of malignant tumors, they are namely Meningioma, Glioma and Pituitary which is classified based on the region in brain where the tumor is present. The existing system uses SVM and KNN combined hybrid classifier (SVM-KNN) for classification of the tumor types. The proposed system uses HOG (Histogram of Oriented Gradients) feature for extraction, AdaBoost for separating tumor and non-tumor feature and Convolutional Neural Network (CNN) to classify the tumor types. The proposed network structure achieves a significant performance with the best overall accuracy for the two methods with accuracy of 98 %. Feature extraction is used to identify the tumor part of the image from the non-tumor part, it is carried out by HOG feature in the program. For classification, CNN is used. Convolutional Neural Network (CNN) is a class of Deep Neural Network and commonly used in analyzing visual imagery. It is inspired by biological processes in human brain and utilized to handle data that come in multiple arrays. The architecture of the network is evolved using different configuration to acquire the most appropriate structure. In CNN architecture, classification of features happens through convolving small filters with the input patterns followed by selection of the most distinguishable feature and then start to train the classification network. The main advantages of CNNs are feature learning and providing more accuracy more than traditional machine learning (SVM-KNN).

| Download

conservation of great white sharks -prevent sharks from extinction

¹Mr.Dhinesh G V , ²Praveenkumar R A

¹Assistant Professor, Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

To prevent the extinction of Great White Sharks using AI, Motion Sensors, An autonomous underwater vehicle

| Download

bitcoin price prediction using recurrence neural network

¹Ms.Hemachitra B , ²Gayathiri B , ³Shivisthika S , ⁴Mrs. Jena Catherine Bel D

¹Assistant Professor, Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Assistant Professor, Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

In this paper, we tried to predict the daily high and closing price, of the Bitcoin. Bitcoins are digital currencies that have gathered noteworthy investor attention in the financial markets. Since, this market is highly volatile and has price fluctuation it needs effective prediction. We attempt to predict the Bitcoin price for the next day by taking various parameters that affect the Bitcoin value such as datetimes, last, diff_24h, diff_per_24h, bid, ask, low, high, volume). Machine learning algorithms like Recurrent Neural Network (RNN) and Long Short- Term Memory (LSTM) have outperformed the traditional time series models in cryptocurrencies price prediction. LSTM basically uses special gates to allow each LSTM layer to take information from both previous layers and the current layer. The data runs through multiple gates (like forget gate, input gate etc.) and activation functions (like tanh function) and is passed through the LSTM cells. The main advantage of this is that it allows each LSTM cell to remember patterns for a certain amount of time they essentially can “remember” important information and “forget” irrelevant information

| Download

authentication using encrypted negative password for online university question papers

¹Mr.Mani Bharathi R , ²Sree Nivetha S , ³Vedhaa Varshini M , ⁴Dr. Chakaravarti S

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Professor , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

In the University exam question papers uploading process, Security is one of the major threats in today’s world. Authorized users should only access the information and no combined algorithm implements hashing, negative password algorithm, and encryption. In existence, the user credentials are stored in a hashed format. Rainbow tables are used to find the plain text corresponding to the hashed values. As the computational speed of the system increases, the rate at which the hacker finds the plain password easily. So two- factor authentication with classical encryption algorithms is not suitable for some areas. So we propose three-factor authentication with Encrypted Negative Password. In this paper, we propose an Encrypted Negative Password to store the user credentials. In this, the user credentials are encrypted and stored. The plain password is converted into a negative password using the Prefix algorithm with permutation to obtain a negative password. Using the symmetric-key algorithm, then the negative password is converted into encrypted Negative Password to improve further security. Using the BLAKE2 algorithm, plain password is hashed to obtain a hashed password. Even if the database is hacked, the password cannot be retrieved as we are using Encrypted Negative Password. The question papers are encrypted using the AES algorithm. When the question paper’s key is shared, it is encrypted using Elliptic Curve Cryptography.

| Download

attendance marking system through face recognition using haar cascade

¹Mr.Annamalai S , ²Arun Kumar M , ³Deeban Balaji J , ⁴Mrs. Jena Catherine Bel D

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

Attendance management system is necessary tool for taking attendance in any environment where attendance is mandatory. However, most of the existing systems are time consuming and it requires manual work from the user. The proposed system is aimed at developing a less time consuming, cost effective and more efficient attendance management system. The significant part of the system consists of database creation, face recognition and attendance mailing. The proposed system is implemented with Haar cascade algorithm which is a object detection algorithm used to identify objects in an image or video. In the proposed system, Haar cascade algorithm is used to identify faces of the people. Images of individuals in various angles will be captured and stored in a database. Image captured by the camera will be processed to find the existence of face. If a face exists then the face is localized and it will be compared with images in the database and returns the individual’s identity. Attendance report will be generated at regular interval as per the requirement and mailed to the person concern.

| Download

arm gesture recognition using accelerometer sensor adxl335 for upper limb amputees

¹Mrs.Revathi S , ²Monika S , ³Nikethana N N , ⁴Noorjahan Reeza A

Abstarct [+]

In today’s lifestyle, technology has become more predominant than it was. A man’s life has become very integrated with the advancements in science and technology that he continues to pursue discovering the depths of both. This project solely focuses on the differently-abled (amputees) who have lost their hands till their wrists or till their forearms. They face a number of problems each day and the supreme one being, staying updated with the technology. Using a computer, laptop, iPad is a challenge because it involves the usage of mouse to browse through the pages. Main objective is to make the usage of the systems for the amputees easier. In this project, we give an insight as to how we built a system from a different perspective which aided us to produce a useful solution for the amputees. This project applies the technology of a microcontroller Arduino for the movement of mouse pointers accordingly. Also, this project uses the ADXL335 component that detects how a user moves that component in various aspects like direction, speed and position. Finally, the project collaborates all these together to form one successful product that can be used as an alias to a hardware mouse.

| Download

age estimation with magnetic resonance imaging data using probabilistic neural network

¹Mr.Abijith S , ²Prem Kumar B , ³Siddarth G , ⁴Ms. Amirthavalli R

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

This paper aims to create a system which provides the age range of a human using a dental MRI image. Age estimation of living individuals or human remains is a very active research field in medicine and forensic application. Age related processes show great variation, both within and between populations. However, in forensic contexts, this parameter is crucial for identification, and both accuracy and reliability are also required. A popular method for forensic age estimation is aspartic acid racemization. The aspartic acid in human tooth enamel shows increasing racemization with age. In this proposed method, age is estimated using Magnetic Resonance Imaging (MRI) data. Discrete wavelet transform is commonly used compression algorithm that utilizes digital image signals. The visual content of the obtained image is captured using Gray-Level Co-occurrence Matrix. A probabilistic neural network finally classifies the images based on extracted features and trained data set to provide the age.

| Download

collaborative access control in corporate system using ip webcam

¹Ms.Pritha M , ²Anupama R , ³Arockia Abins A

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Faculty, Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

Security is one of the major threat in today’s world. Need for Information Security has risen in the view of reducing the risk of unauthorised information disclosure , modification and destruction . In Corporate system, team members request for access to documentation or code repository of projects. Once the requestor is identified to be an authorized user by higher officials, they will be given access rights to the repository.Since this is a one time process, there are high chances for the employee to misuse the company files for the commercial purpose in due course. To overcome this we have proposed a double encryption and IP Webcam authentication technique. The Source code(sensitive data) to be protected is first encrypted using AES algorithm. The key that is used for encrypting the data in AES algorithm is in turn encrypted using the SHA-256 algorithm. To ensure that the requestor is an authorized person, IP WebCam technique is used. Team member has to raise the request to the Team lead. Team lead will accept the request if the member is identified to be an authorized person. In turn the request is sent to the team manager. Once the request is accepted , the team member can download the document. If any of the higher officials reject the request , team member cannot download the file.

| Download

cellular automata based authentication protocol for iot security

¹Mr.Arun Kumar K A , ²Boopalan M , ³Gowthamapriyan A , ⁴Mrs. Hemalatha P

¹Student , Computer Science and Engineering, Adhiyamaan College of Engineering, Hosur, India, India.
²Student , Computer Science and Engineering, Adhiyamaan College of Engineering, Hosur, India, India.
³Student , Computer Science and Engineering, Adhiyamaan College of Engineering, Hosur, India, India.
⁴Assistant Professor, Computer Science and Engineering, Adhiyamaan College of Engineering, Hosur, India, India.

Abstarct [+]

In recent times advancement in information and communication handling technologies made huge impact on Sensing devices. Its lead to all type of devices being connected to the Internet via any connecting protocols like Bluetooth, WLAN, 3G, 4G. Also in recent times cyber-attacks also Increased lot on internet based devices. This will intimate that security schemes under IOT were not fair in concern. Virtually any device connected with internet could be vulnerable to attacking attempts. This paper proposes solution based on the cellular automata based scheme for improving security of the network. In this Each Device having the unique id based on their properties (or random number with time stamp), and while initializing the communication, they will broadcast their id to all neighbour nodes, to pair with other nodes they should exchange their unique id with each other , to send a message between those two nodes we are using cellular automata rule 30 to generate K (key) based on the unique id of the nodes, then we have to store the K. once the receiver received the message means it will generate the K key by itself, if Both the Keys are matching means the Device should be verified , then we have to apply the same cellular automata rule for K (previous key) to authenticate up to last packet received. Also we are applying Link stability approach with rss Algorithm to eliminate the Packet Drop Attacks of the Malicious Nodes. This will results that Security scheme will increase by 30% Expected.

| Download

recoginition of currency for visually challenged people using convolutional neural network

¹Ms.Vishali T , ²Rahhni M , ³Lavanya M , ⁴Dr. Jayalakshmi S L

Abstarct [+]

Visually impaired people face many difficulties in recognizing the currency. Existing systems used content-based methods such as color methods which have the limitations in recognizing the old notes and noisy notes. The Feature extraction methods such as Histogram Oriented Gradients (HOG), Local Binary Pattern (LBP) used by the existing system has difficult computation. Currency recognition using Ensemble Neural Network is done by training each individual neural network. The Proposed approach is based on SIFT (Scale-Invariant Feature Transform) feature extraction and CNN (Convolutional Neural Network) classification. Enhanced Pre-processing techniques such as Gaussian Blurring, Bilateral Filtering and Gray scale conversion is done. The Gaussian Blurring which is the result of blurring an image by a Gaussian function and Bilateral Filtering is another technique which keeps the edges sharp. Gray conversion converts the bilateral filtered image into black and white or gray scale. The enhanced input is passed to the SIFT algorithm for feature extraction which detects and describes the features in images. The extracted features are given to Convolutional Neural Network classifier for classification. After classification, the currency will be analyzed and produced as a speech output with a particular category.

| Download

prevention of deception in bankruptcy using convolutional neural networks

¹Ms.Priyadharshini V , ²Mrs. Sumathi G , ³Vidhya Lakshmi

Abstarct [+]

Bankruptcy is a process through which people seeks relief from the debts which cannot be payed to the creditors. Artificial Intelligence and Machine Learning plays the vital role in the prediction of bankruptcy companies. This paper proposes training algorithm such as Convolutional Neural Network has proven to be efficient in the bankruptcy prediction. The datasets are used to train and test the results using traditional predictionmodel.

| Download

predicting the endurance of malaria by harnessing multilayer perceptron using adam optimizer

¹Ms. Devi K , ²Aishwarya K , ³Varsha T

Abstarct [+]

With the advancement of technology inclination of substituting human efforts in medical field has been flourishing. From the immense knowledge collected, currently the way to analyze the information to predict a patient's sickness and conduct early intervention has become a centered analysis stream. The patient’s intuitive expression of feelings is additionally considered crucial. Doctors record the pathological characteristics of patients within the system. In this paper, we projected the knowledge for the endurance of malaria which is extracted from the multi layer perceptron using the Adam optimizer. A multilayer perceptron may be a category of feed forward artificial neural network.MLP utilizes a supervised learning technique known as back propagation for training. Its numerous layers and non-linear activation distinguish MLP from a linear perceptron. It will distinguish data that's not linearly divisible. Doctors record the pathological characteristics of the patients within the system with a questioner and this knowledge is later examined and checked for the adherence of patient. Adam optimizer has adaptive learning rate, which implies, it computes individual learning rates for various parameters. The Adam optimizer is exerted to extend the optimization. Adherence test is undergone for understanding the drug compliance. Finally cross entropy is performed on the complete model and an accuracy of 90.3% is achieved.

| Download

phishing -malicious email detection using naïve bayesian classifier in data mining

¹Mr.Balaji M , ²Padmanabhan V , ³Tharun Kumar J , ⁴Mr. Arockia Abins A

Abstarct [+]

Phishing is a Cyber Attack which the attacker sends fake emails to attract users to visit fake websites to obtain the user’s personal information. Targeted malicious emails (TME) breaching computer network have become more insidious and more widely documented in recent years. Beyond spam or phishing designed to trick users into revealing personal information, TME can exploit computer networks and gather sensitive information. They can consist of persistent and coordinated campaigns that can span years. A new email-filtering technique based on email's persistent-threat and recipient-oriented features with a Naïve Bayes classifier Algorithm. This paper, how to detect a targeted malicious packet (email) for normal network into modern network. We develop a router detection protocol that dynamically infers the precise number of congestive packet losses that will occur.

| Download

ontology based web search for data mining using stemming algorithm

¹Mr.Rohinth M , ²Rakul C , ³Puspaprem P , ⁴Mr Ganesan R

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

To style a personalized-Suggestion system, a system that makes use of representations of things and user profiles supported ontologies therefore on offer linguistics application with customized services. The recommemder system uses domain ontologies to bolster the personalization. On the one hand, user’s interests area unit modelled throughout an easier and correct means by applying a domain-based logical thinking method; on the other hand, the stemmer formula utilized by our content-based filterate approach, that provides a live of the association between associate degree item and a user, is increased by applying a linguistics similarity methodology. We have a tendency to propose associate degree economical recommender system supported metaphysics and internet Usage Mining. The first step of the approach is extracting options from internet documents and constructing relevant ideas. Then build metaphysics for the web website use the ideas and important terms extracted from documents as per the linguistics similarity of internet documents to cluster then into totally differet linguistics themes, the assorted themes imply totally different preferences. The planned approach integrates well fashioned information into internet Usage Mining and personalization processes.

| Download

online voting system using blockchain technology

¹Ms.Deepa S , ²Sudharshan , ³Jayaraman D , ⁴Raja M

Abstarct [+]

The voting process for a national, state or even selecting a local candidate leader is a crucial and complicated process. This becomes more challenging to a democratic country and for India with more than 1.3 billion people, it is nearly impossible to conduct the voting process. In order to lessen the burden, the e-voting system was introduced in the field but the traditional system has lots of flaws specifically privacy and security issues. Added to that, mistrust in the voting system. In recent years, Blockchain Technology is being used in many fields and has brought significant amount of security and privacy to various processes notably the online transaction. We proposed the usage of Blockchain Technology in e-voting system in order to overcome the flaws and challenges in the existing system. The Blockchain uses decentralized and various authentication process such as Biometric Authentication (facial recognition), security questions and OTP in order to access the voting page. The data is secured using 3-level encryption system and stored in a decentralized system. © 2020 VDGOOD Professional Association. All rights reserved

| Download

location based analysis of women’s safety by using sentiment analysis in social media

¹Mr.Ashwin kumar S , ²Dr. Ramyadevi R , ³Lokeshprabhu U , ⁴Vishal B R

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Assistant Professor, Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

Sexual harassment and other forms of sexual violence in public spaces are an everyday occurrence for women and girls around the world—in urban and rural areas, in developed and developing countries. Rather than imposing restrictions on women that society usually imposes it is the duty of society to imprecise the need of protection of women. As People communicate and share their opinion actively on social medias including Facebook and Twitter, social network can be considered as a perfect platform to learn about people’s opinion and sentiments regarding different events. Analysis of twitter texts collection also includes the name of people and name of women who stand up against sexual harassment and unethical behavior of men in Indian cities which make them uncomfortable to walk freely. The data set that was obtained through Twitter about the status of women safety in Indian society was processed through machine learning algorithms for the purpose of smoothening the data by removing zero values.Since Twitter contains short text, people tend to use different words and abbreviations. These phrases are difficult to extract their sentiment by current NLP systems easily. TextBlob is a high level library that is used to remove links, special characters, etc. from the tweet using some simple regex and tokenize the tweet ,i.e. split words from body of text that is specifically attuned to sentiments expressed in social media. It uses a combination of a sentiment lexicon is a list of lexical features and also tells about how positive or negative a sentiment is. The sentiment analysis process returns the polarity of the tweets (positive/negative).Based on the polarity, a region is marked as safer or not safer.

| Download

malware detection in android application using support vector machine

¹Ms.Poojethaa K , ²Subasree S , ³Nandhinee B , ⁴Rajeswari A

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

The increase in popularity of Android based smart phones attracted many users which in turn lead to the widespread distribution of malicious applications by attackers which resulted in need for sophisticated malware detection techniques. The machine learning classifiers are widely used to model Android malware patterns based on their static and dynamic behaviour. To address the problem of malicious application detection, in this paper we have proposed a machine learning based detection for Android platforms. Our proposed system utilizes the features of collected samples based on classifier to train the apps. The system extracts requested permission which is missing in previous proposed solutions in order to detect the newly invaded malwares in Android platform. To validate the performance of proposed system, various experiments have been carried out to show high accuracy in classification.

| Download

load balancing in cloud using genetic algorithm

¹Dr.Usha M , ²Bharath Kumar B , ³Suresh Kumar M , ⁴Vijay Manikandan R

¹Assistant Professor,, Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

In recent times, cloud computing has become the most powerful and emerging technology. Load Balancing research in cloud technology is one of the lightning technologies in modern time. In pointing to the various proposed algorithms, the topic of load balancing in cloud computing are researched by a gist of the latest ways. Without load balancing load is too much hard to handle. By using Genetic Algorithm the balance is most flexible that is to be represented with this paper.

| Download

lightweight identity-based secured data sharing for university portal

¹Ms.Srutilaya S , ²Srutilaya S , ³T Subashini

Abstarct [+]

A cloud platform provides users with shared data storage services to substantiate shared data integrity, it's a necessity to validate the data effectively. By introducing Hashgraph technology and developing with a third Party Medium (TPM) management strategy, a light- weight secure auditing plan for shared data in cloud storage (LSSA) is projected. Using cloud storage facilities, customers can store their information inside the cloud to prevent disbursement on storing and maintaining native information. Many data integrity audit plans are counseled to confirm the integrity of the data hold on inside the cloud. In our existing work, users ought to use their personal key to create information authenticators to perform data integrity auditing. The user ought to thus have a hardware token (e.g. USB token, intelligent card) to store his personal key and study a watchword for this personal key to be activated. If this hardware token is lost or this watchword is forgotten, most of this auditing systems for data integrity could not operate. In our projected work, a recent paradigm called non-private main storage data integrity audit to prevent practice the hardware token, we tend to tend to use biometric data (e.g.fingerprint) as a result of the user's blurred personal key , exclusively the owner and thus the user is authorized by the administrator to login but it's distinct inside the user module if the user's fingerprint matches that he / she's going to login

| Download

iot enabled bike to improve the road safety

¹Ms.Preethi Bala , ²Nirmal , ³Vinoth R

¹Student , Computer Science and Engineering, Gojan School of Business and Technology, Redhills, India.
²Student , Computer Science and Engineering, Gojan School of Business and Technology, Redhills, India.
³Student , Computer Science and Engineering, Gojan School of Business and Technology, Redhills, India.

Abstarct [+]

Intelligence applications are being developed that make machines more sophisticated in their way of learning and to make decisions. Accident is a specific, unpredicted external action that happens unexpectedly with no apparent or deliberate cause but with marked effects. With the increasing number of bike riders and the number of accidents happening each year our paper focuses on the methods that can be implemented to ensure safety while driving. Distraction of the driver’s attention is the major cause of these accidents. Nowadays wearing helmet has been made mandatory. But still the rules are being violated. Message transmitting sensors are equipped in the speedometer of bike and also in the bike’s helmet. The most important feature of the bike is that the bike’s engine gets start only when the person wears helmet. This system also checks the approaching vehicle’s speed on either side of the road and generate vibrations in the bike’s handlebar. This advanced development is bringing about a new era of productivity for the latest ideas on an astounding scale, understanding their efficiency, speed and functionality. © 2020VDGOODProfessional Association. All rights reserved

| Download

insurance estimation in different stages of chronic diseases using decision tree algorithm

¹Ms.Selshia Kamala Kani I , ²Sujitha E , ³Sushma J , ⁴Mrs. Subashini T

Abstarct [+]

Health insurance fraudulent are detected the most in the recent times especially in case of chronic illness. The chronic disease has its occurrence over a year or two or throughout the lifetime. So providing insurance for the chronic disease is a major issue by considering its stages and symptoms. In the existing model the insurance fraudulent for chronic disease is detected based on the results of Decision Tree algorithm. The Decision Tree is developed with symptoms and stages as nodes .The best attribute of the data-set is kept at the root of the tree. The tree predicts every disease’s stages and the symptoms at each of its stage by analyzing the entire data-set every time. This increases the processing time. In the proposed methodology the Euclidean distance is calculated from the K-Nearest Neighbour algorithm. The training sets that have the merely near Euclidean distance are retrieved and the Decision Tree algorithm is performed only for those sets. The model compares different symptoms associated with different stages of chronic illness. The K-Nearest Neighbour algorithm is used for clustering the chronic diseases with highly related symptoms at different stages by comparing the distance between the query-instance and all the training samples .The searching range for number of symptoms is reduced. In addition the insurance claim is addressed for chronic disease. The insurance claim for the chronic disease is estimated using the previous year records that are claimed for the retrieved stages of the chronic diseases.

| Download

insurance dispute resolution using naive bayes classifier

¹Mr.Ravikumar B , ²Mr. Ravikumar B , ³Harish M , ⁴Jagadeeshwar A

¹Student , Computer Science and Engineering, Velammal Engineering College, Surapet, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

Automobiles are very important to go to workplaces, and to deliver goods. Mostly they lead the way for major problems. Road accident is most worst thing to happen to the people who use roads. The common problem faced by using paper form is the difficulties in retrieving the report back for analyzing purpose as this can be a time consuming. In existing system, a lot of efforts have been earlier done on web based information system in case of road accidents, traffic information, management, analysis and reporting etc. The system is prone to large number of fake claims because there is no proper system to verify if an accident occurred is a real accident or just fake. The information collected from various sources like photos, interview should be used in a efficient manner by sharing them with concerned authorities. Naive Bayes classifier has strong independence assumption between the features. This method is used to categorize text and to judge documents whether it belongs to the same category or not using word frequencies. Also, support vector machines which are supervised learning models with associated learning algorithm that analyse the collected information used for classification and regression analysis. In the same system, there should also be a provision to submit or exchange insurance numbers details in order to settle in order to settle the dispute if any arising out of accident. It encourages prompt reporting and to provide quick action for insurance dispute resolution.

| Download

implementation of traffic light signal using haar cascade algorithm

¹Mr.Bala Suriya S , ²Prithivirajan T , ³Vijayaragavan K , ⁴Dr. Visu P

¹Student , Department of Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

Traffic monitoring and controlling has always been a challenge. In fact the ever increasing day-by-day nature makes it difficult to find where the traffic density is more in real time, to schedule a better traffic signal control and effective traffic routing. The root cause of this can be of different situations like congestion in traffic like insufficient Road conditions due to weather, Road width, large delay of Red Light etc. The solution we provide for Traffic management is that by having a special intelligence which the images of road feed from the cameras (PC Camera) at traffic junctions for the real time traffic density calculation using image processing. It also focuses at reducing the traffic congestion on roads which will help lower the number of accidents. The vehicles will be detected by the system through images instead of using electronic sensors embedded in the pavement. A camera will be placed along the traffic light. It will capture images sequences. Image processing is a better technique to control the state change of the traffic light..Hence it is done by Haar Cascade Algorithm. It shows that this can decrease the traffic congestion and avoids the time wasted by a green light on an empty road.

| Download

optimization of drinking water distribution with flow control system using arduino

¹Mrs.Ramya N , ²Hilary Bernadine J , ³Dharani Devi S , ⁴Kanimozhi R , ⁵Ashasri S

¹Assistant Professor, Department of Electronics and Communication Engineering, SRM-TRP Engineering College, India.
²Student , Electronics and Communication Engineering, SRM-TRP Engineering College,, India.
³Student , Electronics and Communication Engineering, SRM-TRP Engineering College,, India.
⁴Student , Electronics and Communication Engineering, SRM-TRP Engineering College,, India.
⁵Student , Electronics and Communication Engineering, SRM-TRP Engineering College,, India.

Abstarct [+]

The main aim of this paper is to distribute required amount of water from a centralised tank to various distribution points at homes. Water is the most precious resource in the world. In this paper, we could easily detect the leaks in the pipeline using level sensor. The flow sensor generated series of electric pulse through which water utilized by the user, flow rate and the amount of water supplied can be calculated. pH value of water is also monitored using pH sensor to measure the difference in hydrogen ion concentration. Only if the pH value of water in the master tank, is within the specified range, the water is safe and the distribution system is enabled. This paper provides continuous tracking of water usage resulting in avoidance of water wastage. The added outcome of the paper is reduction of manpower and hence cost of the system.

| Download

lifeguard and drowsy detection for drivers

¹Ms.Deepika R S , ²Dr. Akilandeswari , ³Kannika M

¹Student , Computer Science and Engineering, St Joseph’s Institute of Technology,, India.
²Assistant Professor, Computer Science and Engineering, St Joseph’s Institute of Technology,, India.
³Student , Computer Science and Engineering, St Joseph’s Institute of Technology,, India.

Abstarct [+]

We are developing an embedded product that helps in prevention of accidents. Our product has concentrated on two aspects to prevent accidents. First, Drowsiness of drivers. Second, Narrow curves on roads of hills. In order to prevent accidents because of these two aspects our product has been proposed. For preventing accidents because of drowsiness we have implemented image processing based on Eye Aspect Ratio (EAR) [1], it monitors driver and alert them. For preventing at curves we have implemented camera that traps the radius of curve and the vehicle coming opposite. This system also comprises MEMS sensor, GPS, GSM to send message in case of emergency.

| Download

diagnosing the pneumonia disease in secured health records using convolutional neural networks

¹Mr.Navaneetha Krishnan P , ²Mr. Abishek S , ³Mr. Vimal Raj A S

Abstarct [+]

The healthcare sector is very vast and contains data that are sensitive in nature. Every year, there are several million data breaches in the hospitals. This is due to the inefficient encryption techniques. The Public Encryption techniques like SHA and RSA are still in use for securing records. It has authentication problem. Traditional steganography methods hide the secret data by establishing a mapping relationship between the secret data and a cover image which has a low embedding capacity. By using the U-Net Architecture, the image segmentation process is performed. It performs the process of convolution and max pooling, repeatedly in each layer. The secret image (X-Ray) is hidden inside a cover image to produce a container image. By further using a secret key to decrypt the records, the authentication problem is resolved. The detection of pneumonia disease is done by using another deep-learning algorithm called Deep Convolutional Neural Network. The network is processed using Depth-wise Separable Convolution. They are used to learn the features and classify the data based on the image frames. The main advantage of Depth-wise Separable Convolution is that they are computationally cheaper i.e. they require fewer computations than standard technique

| Download

obstacle detection for visually challenged using surf

¹Ms.Monisha S , ²Saranya M , ³Sulochana V , ⁴Hemalatha B

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student, Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

According to WHO, it is estimated that approximately 1.3 billion people live with some form of vision impairment where India is being the home of world’s largest number of visually challenged people. Visually challenged perform everyday tasks with certain amount of restrictions in mobility due to obstacles in path. To improve the quality of life many embedded systems exists to assist them for obstacle detection using techniques like Stereo Vision, Scale- Invariant Feature Transform(SIFT) and Fast library for approximate Nearest Neighbors. Further supports in mobility to visually impaired people is proposed using image processing technique in obstacle detection to provide obstacle notification with voice assistant. The detection of obstacle is implemented through a KNN classifier and feature extraction is done by SURF algorithm. SURF extracts features of obstacle images captured from different perspectives to estimate segmental key points. These key points are used for categorizing object with the help of KNN classifier. The detected object is intimated through voice assistant to visually impaired people that obstruct their path, thereby warning them for safe mobility.

| Download

a survey on cancer detection using machine learning algorithms

¹Mr.Anandatirtha J S , ²Dr Janarthanan P , ³Deepak S

¹Student, Computer Science and Engineering, Sri Venkateswara Colle ge of Engineering, , India.
²Associate professor, Computer Science and Engineering, Sri Venkateswara Colle ge of Engineering, India.
³Student, Computer Science and Engineering, Sri Venkateswara Colle ge of Engineering, India.

Abstarct [+]

Cancer is a major disease characterized by the development of abnormal cells that divide uncontrollably and destroy the normal body tissues. When the possibility of cancer is determined and the treatment can be started earlier in the course of disease possibly before disease spreads. Machine Learning has seen tremendous growth to diagnose various diseases and the algorithms can handle multi-dimensional and multi-variety data which perform this in dynamic or uncertain environments. The algorithms like Convolution Neural Networks (CNN), Naïve Base classifier (NB), Decision-Tree (DT), Fuzzy clustering algorithms and other techniques are used to diagnose cancer. Even though it is evident, there is still space in diagnosing cancer at an earlier stage. We confer research work to analyze the various algorithms utilized in cancer research

| Download

classification and segmentation of kidney stone and cyst using otsu segmentation algorithm

¹Mr.Bhagyasree Atlaa , ²Iswarya M , ³Sujalakshmi S , ⁴Bhagyalakshmi A

¹Student, Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Assistant Professor, Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

Due to high grey intensity of the MRI image, the kidney stone and cyst goes unnoticed by human inspection, this leads to inaccuracies in detection due to noise in MRI images. Detecting kidney stone and cyst using Back Propagation Neural Network (BPNN) calculates weight based on features extracted through Gray Level Co-occurrence Matrix (GLCM) for every images. K-means clustering does only the distance calculation to detect the kidney stone and cyst region. Thus it is slow and time inefficient. In our proposed system, it preprocesses the MRI image and uses Discrete Wavelet Transformation (DWT) algorithm for MRI image enhancement. GLCM is used for feature extraction, since GLCM acquires features by considering neighbor pixels thus all minute information are considered, thus it is very efficient in feature extraction. Support Vector Machine (SVM) is used to detect whether it is a kidney stone or cyst. Otsu segmentation algorithm returns a gray level intensity value that inturn used for the separation of the pixels into two classes such as foreground and background. Foreground region identifies the kidney stone or cyst region. Based on the features extracted, we determine the size and location of the kidney stone or cyst. The Analysis process is done based on the intensity threshold variation obtained from the segmented portions of the image and size of the portions compared to that of the standard stone sizes.

| Download

computer vision based attendance monitoring system

¹Mr.Kamaleshwaran K , ²Jushwin Rassa J , ³Vigneshwar V , ⁴A.Moorthy

¹UG students, , Computer Science and Engineering, St.Peter’s College of Engineering and Technology, India.
²Student , Computer Science and Engineering, St. Peter’s College of Engineering and Technology, India.
³Student , Computer Science and Engineering, St. Peter’s College of Engineering and Technology, India.
⁴Student , Computer Science and Engineering, St. Peter’s College of Engineering and Technology, India.

Abstarct [+]

Today there is a need for automation systems and for this we can automate on how the attendance is done. The face detection system has also delivered a security enhancement, where this can be used in the system to provide greater security. There is no need of manual interventions and the attendance is taken automatically and is stored in the database.

| Download

smart crop irrigation system using iot and detection of leaf diseases

¹Dr.Vinoth G , ²Mr. Vignesh K , ³Mr. Vijay P , ⁴Mr. Lingesh Kumar C

¹Assistant Professor, Electrical and Electronics Engineering, Adithya Institute of Technology, India.
²Researcher, Electrical and Electronics Engineering, Adithya Institute of Technology, India.
³Researcher, Electrical and Electronics Engineering, Adithya Institute of Technology, India.
⁴Researcher, Electrical and Electronics Engineering, Adithya Institute of Technology, India.

Abstarct [+]

Agriculture plays a vital role in the development of agricultural country like India. Issues concerning agriculture have been always hindering the development of the country like India. The only solution of this problem is smart agriculture by modernizing the current traditional methods of agriculture using IoT. Hence the proposed ideas to making the agriculture smart using automation and used to detect the plant leaf diseases using IoT technologies. Internet of Things (IoT) enables various applications like crop growth monitoring and selection, irrigation decision support, also used for fertilizer, analysis the report of detected disease etc. A Raspberry Pi is based on automatic irrigation IoT system is proposed to modernization and improves productivity of the crop and growth, main aim of this work to crop development on low quantity water consumption, in order to focus the water available to the plants at the required time for concern level, for that purpose most of the farmers waste lot time in the fields. An efficient management of water should be developed and the system circuit complexity to be reduced. The proposed system developed the information sent from the sensors and estimate the quantity of water needed for the soil. The objective of this work is to implement image analysis & classification techniques for detection of leaf diseases. Two sensors are used to get the data to the base station the humidity and the temperature of the soil, duration the sunshine per day. The proposed systems based on those values and calculate the water quantity for irrigation is required. The major advantage the system is implementing of Precision Agriculture (PA) with cloud computing, that will optimize the water scarcity.

| Download

aid to the visually challenged tutor using convolutional neural network

¹Dr.Usha Ma , ²Greeshma M , ³Latha Sri S , ⁴Shalini N

Abstarct [+]

Visually Challenged Tutors cannot visualize the students reaction. The only way they could understand is, by the oral feedback to the teachers. The Facial expressions are one of the most convenient way of identifying the student's state of mind. The existing methods for facial expressions uses Support Vector Machine (SVM) method. In order to avoid the complex process of explicit feature extraction and low level data manipulation, a fast CNN facial expression recognition method is used. The aim of our proposed system is to detect the emotion of the students from facial expression using Convolutional Neutral Network (CNN). HAAR Cascade Classifier algorithm is used to detect the facial features of the students. The edge of each layer of the image is extracted in the convolution process. The edge information is extracted and is superimposed on each feature image to identify the edge structure information of the texture image. The implicit feature is extracted and is processed by maximum pooling method. Finally the expression of the test sample image is identified. By taking the mean of the students emotion and comparing it with a threshold value, the ambience of the class room is determined. A comment regarding the atmosphere of the classroom is conveyed to the visually impaired teacher through an audio device periodically. Additionally, an analysis system for teachers based on maintaining an active environment for students in the classroom is developed and suggestions for improvisation will be given.

| Download

detection of money laundering using web application

¹Prof.Subramaniam .M , ²Ajayraj.M , ³Kalanidhi.G , ⁴Santhosh.V

¹Professor, Department of Computer Science and Engineering, S.A. Engineering College, India.
²Researcher, Department of Computer Science and Engineering, S.A engineering college, India.
³Researcher, Department of Computer Science and Engineering, S.A engineering college, India.
⁴Researcher, Department of Computer Science and Engineering, S.A engineering college, India.

Abstarct [+]

Misrepresentation discovery systems for national and global economies have progressed toward becoming a significant vital undertaking. Guaranteeing the security of exchanges did by banks and other money related establishments is one of the central point influencing the notoriety and gainfulness of such associations. Be that as it may, since individuals who perform fake exchanges change their techniques continually all together not to get made up for lost time, it gets harder to distinguish and identify this sort of exchanges. Identifying this sort of exchanges makes the help of innovation necessary, considering high volume and power of exchanges. In this paper, we investigate reasonableness of utilizing area information to help discovering better business rules where they can without much of a stretch be sent with a control based extortion discovery and avoidance framework for retail keeping money. Contingent upon how much versatile the card proprietors are, we can without much of a stretch devise business tenets to identify the abnormalities. We have appeared in this paper a noteworthy greater part of clients does not leave the region of their living spot. We additionally give some concise utilize cases and indications with respect to what sorts of business tenets can be extricated from area information.

| Download

smart stretcher and integrated medical intelligence system for unconscious person

¹Ms.Sowmiya J , ²Malathi K , ³Ashwini L , ⁴S.Mary Cynthia

¹Researcher, Electronics and Communication Engineering , Jeppiaar Institute of Technology, India.
²Researcher, Electronics and Communication Engineering , Jeppiaar Institute of Technology (Anna University), India.
³Researcher, Electronics and Communication Engineering , Jeppiaar Institute of Technology (Anna University), India.
⁴Assistant Professor, Electronics and Communication Engineering , Jeppiaar Institute of Technology (Anna University), India.

Abstarct [+]

India is one of the most populous countries of the world. Due to overpopulation, ignorance of health has remained a major problem in india. For every one minute a death swoops in because of unpredictable and unexpected accidents. To save a life is auspicious as well as precious. The idea here is to provide an intelligent smart health system using some sensors and microcontrollers which are implemented in stretchers and also a live update of the patient is sent to the nearby hospital and police station through iot. IOT based health monitoring system to analyze and compute the patient health smart devices connected to the internet for communicating with each other. The aim of this system is to save many human lives by preparing intensive care units in hospitals, as their physical parameters are updated to hospital before their arrival to hospital.

| Download

smart navigation using augmented reality

¹Mr.Sai Suresh M V , ²Vaishali S , ³Subashini R , ⁴Sri Charan K V

¹Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
²Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
³Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.
⁴Student , Computer Science and Engineering, Velammal Engineering College, Chennai, India.

Abstarct [+]

Augmented Reality is a combination of a real and a computer-generated or virtual world. It is achieved by augmenting computer-generated virtual objects in the real world. It is of four types namely marker-based, markerless, projection-based and superimposition based augmented reality. AR has many applications in the real world. Augmented reality comes under the field of mixed reality. It can be considered as an inverse reflection of Virtual Reality. This paper gives information about Augmented Reality and how it started. This paper also gives us knowledge regarding its current and future applications on marketing. The proliferation of AR-based indoor navigation in different consumer sectors is estimated to experience tremendous growth in the coming years as technologies continue to evolve. As users increasingly become more digital-savvy and ready to apply new technologies in their everyday life. It gives us a comparison between the two related topics, Augmented reality and Virtual reality. The following paper also helps us know about the effect of Augmented Reality on human life.© 2020VDGOOD Professional Association. All rights reserved

| Download

Sign In

Reset My Password

VDGOOD Journal of Computer Science Engineering

Volume: 1- Issue: 2