Deep Similarity Metric Learning for Real-Time Pedestrian Tracking

TL;DR

This paper introduces a deep similarity metric learning approach for real-time pedestrian tracking that enhances tracking accuracy by learning appearance embeddings, improving ID consistency, occlusion handling, and detection proposals.

Contribution

It presents a novel integration of a deep Siamese network for appearance embedding into a real-time tracking framework, improving multiple aspects of pedestrian tracking.

Findings

Achieves competitive results among online real-time trackers.

Improves ID switch prevention and occlusion handling.

Demonstrates the effectiveness of deep appearance metrics in tracking.

Abstract

Tracking by detection is a common approach to solving the Multiple Object Tracking problem. In this paper we show how learning a deep similarity metric can improve three key aspects of pedestrian tracking on a multiple object tracking benchmark. We train a convolutional neural network to learn an embedding function in a Siamese configuration on a large person re-identification dataset. The offline-trained embedding network is integrated in to the tracking formulation to improve performance while retaining real-time performance. The proposed tracker stores appearance metrics while detections are strong, using this appearance information to: prevent ID switches, associate tracklets through occlusion, and propose new detections where detector confidence is low. This method achieves competitive results in evaluation, especially among online, real-time approaches. We present an ablative study showing the impact of each of the three uses of our deep appearance metric.