Title of the talk:
" 2D TO 3D - MACHINE LEARNING FOR 3D RECONSTRUCTION "
Professor Santanu Chaudhury
Dept. of Electrical Engneering
Indian Institute of Technology, Delhi
Hauz Khas, New Delhi – 110 016.
A fundamental problem in Computer Vision is construction of 3D model based upon 2D images. 'Shape from X' techniques have been developed to address these problems. However, in recent times non-classical techniques using different variants of machine learning has been proposed to improve upon Shape from X techniques for extracting depth information from 2D images. In this talk I shall address these machine learning based techniques. We shall present a new approach based upon manifold learning. By formulating consistent depth estimation as a problem of sequential labeling, we shall illustrate how a new paradigm of fixed-point learning can yield more accurate depth estimation.
Professor Santanu Chaudhury did his B.Tech (1984) in Electronics and Electrical Communication Engg and Ph.D (1989) in Computer Science and Engg. from I.I.T, Kharagpur, India. Currently, he is Director of CEERI Pilani and Professor, on lien, from the department of Electrical Engg. at I.I.T,Delhi. He was also Dean, Undergraduate Studies at IIT Delhi. He was awarded INSA medal for young scientists in 1993. He is a fellow of Indian National Academy of Engineering and National Academy of Sciences, India. He is also fellow of International Association of Pattern Recognition. His research interests are in the areas of Computer Vision, Robotics, Embedded Systems and Machine Learning. He has published more than 250 papers in international journals and conference proceedings. He has been on the programme committee of a number of international conferences like, ICCV, ACCV, ICPR, ICVGIP, PReMI etc.