Speaker: Rob Fergus, NYU

Time: Monday, Feb 9, 2PM
Location: Babbio 221, Stevens Institute of Technology
Host: Philippos Mordohai

Abstract:
The vast majority of information on the Internet is in visual form, yet we currently lack effective methods for searching images or videos. Existing strategies rely mainly on textual cues which give impoverished and often misleading descriptions of the visual content. A key part of the challenge is that the search needs to be highly efficient due to the scale of the problem: Google's Image Search indexes around 10 billion images, while YouTube holds petabytes of video data and receives 10 hours of new content each minute.

In my talk I will describe methods for efficiently searching Internet-sized image databases. Using machine learning techniques, we represent each image with a compact binary code, at most a few hundred *bits* in length, which preserves the original neighborhood structure of images in the database. Our scheme is able to perform real-time search on millions of images using a standard PC, obtaining a retrieval performance comparable with that of more complex descriptors, despite being many orders of magnitude faster.

Joint work with: Antonio Torralba (MIT), Yair Weiss (Hebrew University).

Bio:
Rob Fergus is currently an Assistant Professor of Computer Science at the Courant Institute of Mathematical Sciences, New York University. Originally from the UK, Fergus did his undergraduate degree in Electrical Engineering at the University of Cambridge. He then did a Masters in Electrical Engineering with Prof. Pietro Perona at Caltech, before completing a PhD with Prof. Andrew Zisserman at the University of Oxford. Before coming to NYU, Fergus spent two years as a post-doc in the Computer Science and Artificial Intelligence Lab (CSAIL) at MIT, working with Prof. William Freeman. In 2003, he and his co-authors were awarded the CVPR Best Paper Prize. In 2005, his PhD thesis won the prize for the best Computer Science thesis in the UK.