Brownian descriptor: a Rich Meta-Feature for Appearance Matching

Slawomir Bak 1 Ratnesh Kumar 1 François Bremond 1
1 STARS - Spatio-Temporal Activity Recognition Systems
CRISAM - Inria Sophia Antipolis - Méditerranée
Abstract : This paper introduces an image region descriptor and applies it to the problem of appearance matching. The proposed descriptor can be seen as a natural extension of covariance. Driven by recent studies in mathematical statistics related to Brownian motion, we design the Brownian descriptor. In contrast to the classical covariance descriptor, which measures the degree of linear relationship between features, our novel descriptor measures the degree of all kinds of possible relationships between features. We argue that the proposed covariance is a richer descriptor than the classical covariance, especially when fusing non-linearly dependent features. We demonstrate that the Brownian descriptor significantly outperforms the classical covariance in terms of matching accuracy and efficiency, opening new directions for future research.
Document type :
Conference papers
Complete list of metadatas

https://hal.inria.fr/hal-00905588
Contributor : Slawomir Bak <>
Submitted on : Monday, November 18, 2013 - 2:28:07 PM
Last modification on : Tuesday, July 24, 2018 - 3:48:06 PM

Identifiers

  • HAL Id : hal-00905588, version 1

Collections

Citation

Slawomir Bak, Ratnesh Kumar, François Bremond. Brownian descriptor: a Rich Meta-Feature for Appearance Matching. WACV: Winter Conference on Applications of Computer Vision, Mar 2014, Steamboat Springs CO, United States. ⟨hal-00905588⟩

Share

Metrics

Record views

324