Robust Appearance-Based Object Recognition Using a Fully Connected Markov Random Field

This paper presents a new kernel method for appearance-based object recognition, highly robust to noise and occlusion. It consists of a fully connected Markov Random Field that integrates results of Spin Glass theory with Gibbs probability distributions via nonlinear kernel mapping. We call the resulting model Spin Glass-Markov Random Field. We present theoretical analysis and several experiments that show its effectiveness and robustness to noise and occlusion. We obtain in both cases excellent results. Particularly, we achieve a recognition rate above 93 % with just 40 % of visible portion of the object.

• Type: Conference
• Year: 2002
• Where: ICPR
• Authors: Barbara Caputo, Sahla Bouattour, Heinrich Niemann
• Keywords: Appearance-Based Object Recognition, Computer Vision, ICPR 2002, Markov Random Field, Nonlinear Kernel Mapping