Bags of Affine Subspaces for Robust Object Tracking

TL;DR

This paper introduces an adaptive object tracking method that models the object as a dynamic collection of affine subspaces, improving robustness against occlusions, deformations, and appearance changes by leveraging Grassmann manifold geometry.

Contribution

The novel approach models objects as a bag of affine subspaces and compares them using subspace-to-subspace distances on Grassmann manifolds, enhancing tracking robustness.

Findings

Achieves higher accuracy than recent discriminative trackers.

Effectively handles occlusions, deformations, pose, and illumination changes.

Demonstrates robustness on challenging video sequences.

Abstract

We propose an adaptive tracking algorithm where the object is modelled as a continuously updated bag of affine subspaces, with each subspace constructed from the object's appearance over several consecutive frames. In contrast to linear subspaces, affine subspaces explicitly model the origin of subspaces. Furthermore, instead of using a brittle point-to-subspace distance during the search for the object in a new frame, we propose to use a subspace-to-subspace distance by representing candidate image areas also as affine subspaces. Distances between subspaces are then obtained by exploiting the non-Euclidean geometry of Grassmann manifolds. Experiments on challenging videos (containing object occlusions, deformations, as well as variations in pose and illumination) indicate that the proposed method achieves higher tracking accuracy than several recent discriminative trackers.