I am currently pursuing PhD with Computer Engineering, University of New Mexico. I am being advised by Dr. Marios Pattichis. My areas of emphasis include Image analysis, Video analysis (Human activity recognition), Video compression (HEVC, VP9) and Video delivery (MPEG DASH). In addition I have good experience in Web development, Statistical analysis, and Machine learning. I received my Bachelors degree in ECE (Electrical and Computer Engineering) from VIT (Vellore Institute of Technology). Before joining IVPCL, I have two years industrial experience in Video Compression.
Drastic focuses on developing adaptive video processing system that can respond to changes in content, environment and needs of user. This adaptability is demonstrated using reconfigurable hardware and software specific to HEVC (High Efficiency Video Coding). In this current iteration, I am involved in developeing a codec agnostic frame work that delivers optimal video stream over HTTP using state of the art adaptive delivery mechanism.
AOLME project implements integrated curriculum for teaching computing foundations based on middle-school mathematics. This involves collection of Tera Bytes of videos to develop learning models that help in understanding how students acquire comptuing knowledge. I am involved in developing video analysis methods (human activity recognition, screen activity recognition) that aid in developing learning models.
Images with region of interest (roi) are most common output of forecasting systems which study and map evolution of astronomical events. The models built for such natural events tend to have multiple random free parameters resulting in multiple forecasting maps. Due to inherent randomness, it is impossible to identify optimal free parameters, requiring dynamic classification of models as good or bad. In this paper we attempt to solve one such classification problem by carefully designing and employing protocols to subjectively classify models as good or bad. Upon subjective classification, an automated framework is created to do the same.
“Automatic Segmentation of Coronal Holes in Solar Images and Solar Prediction Map Classification ,” M.Sc. Thesis, ECE, Fall 2016
G. Esakki, V. Jatla and M. S. Pattichis, "Adaptive High Efficiency Video Coding Based on Camera Activity Classification," 2017 Data Compression Conference (DCC), Snowbird, UT, 2017, pp. 438-438.
C. W. Eilar, V. Jatla, M. S. Pattichis, C. LópezLeiva and S. Celedón-Pattichis, "Distributed video analysis for the advancing out of school learning in mathematics and engineering project," 2016 50th Asilomar Conference on Signals, Systems and Computers, Pacific Grove, CA, 2016, pp. 571-575.
G. Esakki, V. Jatla and M. S. Pattichis, "Optimal HEVC encoding based on GOP configurations," 2016 IEEE Southwest Symposium on Image Analysis and Interpretation (SSIAI), Santa Fe, NM, 2016, pp. 25-28.
Pattichis, M.S., Hock, R., Jatla, V., Henney, C., Arge, C., “ Detecting Coronal Holes for Solar Activity Modeling ,” 2014 Asilomar Conference on Signals, Systems, and Computers , 5 pages, 2014.
“System and Methods for Adaptive Optimization for Video Coding and Video Delivery,"
Inventors: Marios S. Pattichis, Yuebing Jiang, Cong Zong, Gangadharan Esakki, Venkatesh Jatla, and Andreas Panayides, Patent filed on 07/31/2015. The patent application describes a novel method to support the optimal delivery of high efficiency video encoded (HEVC) video by adaptive encoding based on content, network bandwidth, or user expectations. It describes an important extension of the awarded U.S. Pattent 9,111,059 B2 on System and Methods for Dynamic Management of Hardware Resources described below. Refer to the architectures research projects and the DRASTIC research project for applications. This material is based upon work supported by the National Science Foundation under NSF AWD CNS-1422031. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. The patent is now available for licensing.