Qualification: 
M.Tech, BE
saraths@am.amrita.edu

Sarath S. currently serves as an Assistant Professor  (Sr. Gr.) at the Department of Computer Science Engineering at Amrita School of Engineering, Amritapuri. He has completed M. Tech. in Computer Science and Engineering from MS University. He has 14 years of academic experience.

Publications

Publication Type: Conference Proceedings

Year of Publication Title

2019

S. Menon, Swathi, J., Anit, S. K., Nair, A. P., and Sarath S., “Driver face recognition and sober drunk classification using thermal images”, Proceedings of the 2019 IEEE International Conference on Communication and Signal Processing, ICCSP 2019. Institute of Electrical and Electronics Engineers Inc., Melmaruvathur; India, pp. 400-404, 2019.[Abstract]


A major share of accidents happening today is categorised under drunk and drive accidents. Attempts to curb these accidents are limited to manual checking of drivers and awareness programs, which is evidently not enough or stringent. We propose a system where the driver's face is captured in thermal image spectrum and is first recognised using facial recognition, then classified as drunk or sober. The former is done using a deep learning tool that is Convolution Neural Network and the latter is done using Gaussian Mixture Model along with Fischer Linear Discriminant for dimensionality reduction. Post the facial recognition, we will be using capillary junction points on faces to determine difference in blood temperature thus allowing us to classify them as drunk or not. © 2019 IEEE.

More »»

2019

A. Ravindran, G. Krishna, G., Sagara,, and Sarath S., “A comparative analysis of machine learning algorithms in detecting deceptive behaviour in humans using thermal images”, Proceedings of the 2019 IEEE International Conference on Communication and Signal Processing, ICCSP 2019. Institute of Electrical and Electronics Engineers Inc., Melmaruvathur; India, pp. 310-314, 2019.[Abstract]


Various researches are going on to identify the deceptive behaviour in humans by means of modern technology. The traditional and long term method used for detection is the polygraph technology, however it is not considered reliable. Impressive research works have been made in the last few decades to detect the deceptional behaviour in humans in a most promising way. Thermal imaging technology which is a rising trend can be applied to figure out the stress levels in humans by assessing the heat radiated from facial region. In this approach we record the difference in temperature in the pre-orbital region to discriminate between a truth teller and a liar. This paper aims at a comparative study of machine learning algorithms on a facial thermal image data set. By using the three algorithms K-Nearest Neighbor, Logistic Regression and Decision Tree we attempted to create a method that can detect lies. We arrived at a solution that gave 90% accuracy for K-Nearest Neighbor, Logistic Regression each and 85% accuracy for Decision Tree. © 2019 IEEE.

More »»

2018

Sarath S. and Anamika, A. M., “Real-Time Video Segmentation Using a Single Click”, Proceedings of the 2018 IEEE International Conference on Communication and Signal Processing, ICCSP 2018. Institute of Electrical and Electronics Engineers Inc., Adhiparasakthi Engineering College, Melmaruvathur, Chennai, pp. 448-451, 2018.[Abstract]


Introducing a method to segment the region of interest ROI from a video file using a single click from the user. The foreground section of the image is hand-picked by the user while the background section of the image is automatically identified by the system. Penalty based random walker algorithm is used for segmentation in this method. The efficiency of the segmentation algorithm is improved by the use of penalty term while calculating the weighting function, thus making this method more user-friendly and effective. © 2018 IEEE.

More »»

2018

Sarath S. and Deepthi, L. R., “Priority Based Real Time Smart Traffic Control System Using Dynamic Background”, 2018 International Conference on Communication and Signal Processing (ICCSP). IEEE, Chennai, India, pp. 620-622, 2018.[Abstract]


Vehicular traffic is increasing rapidly in this world which is resulting in traffic congestion. The emergency vehicles such as ambulances, fire engines and police vehicles are given equal priority over other vehicles and hence get stuck up in this traffic congestion. A methodology for priority based vehicle detection based on image processing techniques is proposed in this paper. If an emergency vehicle is detected on the road, the lane in which this vehicle is will be given higher priority over all other lanes. The paper proposes an algorithm which will detect whether a vehicle is an emergency one or not.

More »»

2016

Sarath S., Chinnu, R., and Gopika, P. S., “Real time smart traffic control system using dynamic background”, International Conference on Computing Paradigms (ICCP 2016) (International Journal of Control Theory and Applications), vol. 9. International Science Press, Serials Publications, pp. 4249-4255, 2016.[Abstract]


Introducing a method for controlling traffic congestion on roads using real time smart traffic control system with dynamic background. This work proposes a real time traffic system in which it uses dynamic background and foreground images for the determination of traffic density based on image processing techniques. It is used for calculating traffic density rather than counting the number of vehicles. The traffic density is calculated by combining gradient magnitude and direct subtraction methods. © International Science Press.

More »»

Publication Type: Journal Article

Year of Publication Title

2015

J. A. James, Karthika, S., Keerthana, S. K. K., Reshma, P. K., Sarath S., Sathyadevan, S., and Hariram, S., “Generalization of crossfilter.Js for amrita dynamic dashboard”, International Journal of Applied Engineering Research, vol. 10, pp. 22818-33815, 2015.[Abstract]


Data Visualization enables data analysts to effectively discover patterns in large multivariate data sets through graphical means and to represent these findings in a meaningful and effective way. The crux of this paper is to ideally incorporate the concept of generalizing the functions pertaining to Crossfilter Application Programming Interface thus efficiently employing the grouping and filtering capabilities of the JavaScript library dc.js (which relies on two other JavaScript plug-in/libraries: D3.js and Crossfilter.js). Consequently the data is visualized as multiple charts like bar chart, pie chart, line chart (available in dc.js library). © Research India Publications.

More »»