A Graph Transduction Game for Multi-target Tracking

TL;DR

This paper introduces a novel graph transduction game approach for multi-target tracking in videos, modeling targets as players in a game to improve tracking accuracy using semi-supervised learning.

Contribution

It applies a game-theoretic graph transduction method to multi-target tracking, demonstrating robustness and effectiveness in semi-supervised video analysis.

Findings

Achieves satisfactory results on surveillance datasets.

Robust to unbalanced labeled and unlabeled data.

Utilizes covariance matrix distances for target similarity.

Abstract

Semi-supervised learning is a popular class of techniques to learn from labeled and unlabeled data. The paper proposes an application of a recently proposed approach of graph transduction that exploits game theoretic notions to the problem of multiple people tracking. Within the proposed framework, targets are considered as players of a multi-player non-cooperative game. The equilibria of the game is considered as a consistent labeling solution and thus an estimation of the target association in the sequence of frames. Patches of persons are extracted from the video frames using a HOG based detector and their similarity is modeled using distances among their covariance matrices. The solution we propose achieves satisfactory results on video surveillance datasets. The experiments show the robustness of the method even with a heavy unbalance between the number of labeled and unlabeled input patches.