A Grassmann Manifold-based Domain Adaptation Approach

Published

November 20, 2012

Author(s)

P J. Phillips, Jingjing Zheng, Ming-Yu Liu, Rama Chellappa

Abstract

Domain adaptation algorithms that handle shifts in the distribution between training and testing data are receiving much attention in computer vision. Recently, a Grassmann manifold-based domain adaptation algorithm that models the domain shift using intermediate subspaces along the geodesic connecting the source and target domains was presented in [6]. We build upon this work and propose replacing the step of concatenating feature projections on a very few sampled inter- mediate subspaces by directly integrating the distance between feature projections along the geodesic. The proposed approach considers all the intermediate sub- spaces along the geodesic. Thus, it is a more principled way of quantifying the cross-domain distance. We present the results of experiments on two standard datasets and show that the proposed algorithm yields favorable performance over previous approaches.

Proceedings Title

Proceedings of the International Conference on Pattern Recognition

Conference Dates

November 11-15, 2012

Conference Location

Tuskuba Science City

Conference Title

International Conference on Pattern Recognition

Pub Type

Conferences

Download Paper

Local Download

Keywords

biometrics

Citation

Phillips, P. , Zheng, J. , Liu, M. and Chellappa, R. (2012), A Grassmann Manifold-based Domain Adaptation Approach, Proceedings of the International Conference on Pattern Recognition, Tuskuba Science City, -1, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=912700 (Accessed October 20, 2025)