I am a Machine Learning and Computer Vision researcher. I successfully defended my Ph.D. thesis at UMass Amherst in Oct 2013.
Because of its unique challenges and huge potential for useful applications, I find computer vision to be an exciting field to apply machine learning in. I have worked on a range of problems including motion segmentation, face detection, face pose estimation, object detection, and classification. I am also interested in applications of machine learning in domains such as text analysis, finance, bioinformatics, and recommendation engines.
My PhD thesis is about motion segmentation in moving as well as stationary cameras. Potential platforms for our moving camera motion segmentation algorithm are smart-phones, automobiles, and mobile robots. Here is our ICCV 2013 paper describing our system [pdf]. Background subtraction researchers may find our clean, complete, and intuitive view of background modeling useful. Here is the journal paper [pdf]. .
I have worked as an intern at Mathworks adding a cascade object detection (boosting) training framework to the Computer Vision Systems Toolbox in Matlab - the language so many machine learning and vision researchers use everyday.
I was also in a start-up vision company that used object detection in video broadcasts to place ads on them.I sometimes tweet about machine learning, vision, and other stuff. Follow @manjunuti