Computational Vision at UC Irvine

Overview:

A large precentage of the images and video that exists today - from personal photo collections to webcams to online images - contain people. We are working on automated methods for analyzing people from such data. Such methods have a variety of applications from pedestrian detection for autonomous vehicles to surveillance-based security to motion capture for computer graphics & interaction. We are also interested in higher-level activity models that can capture human actions and behavior.

Publications:

Ramanan, D.

Using segmentation to verify object hypotheses

Cvpr, 2007, p. 1-8

Allin, S.; Ramanan, D.

Assessment of post stroke functioning using machine vision

Iapr conference on machine vision applications (mva), 2007,

Ramanan, Deva; Baker, Simon; Kakade, Sham

Leveraging archival video for building face datasets

Iccv, 2007, 141,

Ramanan, Deva

Learning to parse images of articulated bodies

Advances in neural information processing systems 19, 2007, p. 1129--1136

Ramanan, D.; Forsyth, D.A.; Zisserman, A.

Tracking people by learning their appearance

PAMI, 2007, 29, 291, p. 65-81

Ramanan, D.; Sminchisescu, C.

Training deformable models for localization

Cvpr, 2006, p. I: 206-213

Ramanan, D.; Forsyth, D.A.; Zisserman, A.

Tracking people and recognizing their activities

Cvpr, 2005, p. II: 1194

Forsyth, David A.; Arikan, Okan; Ikemoto, Leslie; O'Brien, James; Ramanan, Deva

Computational studies of human motion: part 1, tracking and motion synthesis

Found. Trends. Comput. Graph. Vis., 2005, 1, 12-3, p. 77--254

Ramanan, Deva; Forsyth, David A.

automatic annotation of everyday movements

Advances in neural information processing systems 16, 2004,

Ramanan, D.; Forsyth, D.A.

Finding and tracking people from the bottom up

Cvpr, 2003, p. II: 467-474