Title

Training a Scene-Specific Pedestrian Detector using Tracklets

Abstract

A generic pedestrian detector trained from generic datasets cannot solve all the varieties in different scenarios, thus its performance may not be as good as a scene-specific detector. In this paper, we propose a new approach to automatically train scene-specific pedestrian detectors based on track lets (chains of tracked samples). First, a generic pedestrian detector is applied on the specific scene, which also generates many false positives and miss detections, second, we consider multi-pedestrian tracking as a data association problem and link detected samples into track lets, third, track let features are extracted to label track lets into positive, negative and uncertain ones, and uncertain track lets are further labeled by comparing them with the positive and negative pools. By using track lets, we extract more reliable features than individual samples, and those informative uncertain samples around the classification boundaries are well labeled by label propagation within individual track lets and among different track lets. The labeled samples in the specific scene are combined with generic datasets to train scene-specific detectors. We test the proposed approach on three datasets. Our approach outperforms the state-of-the-art scene-specific detector and shows the effectiveness to adapt to specific scenes without human annotations.

Meeting Name

2015 IEEE Winter Conference on Applications of Computer Vision (2015: Jan. 5-9, Waikoloa, HI)

Department(s)

Computer Science

Keywords and Phrases

Computer Vision; Classification Boundary; Data Association Problem; False Positive; Human Annotations; Label Propagation; Miss Detection; Pedestrian Tracking; State of the Art; Object Detection

International Standard Book Number (ISBN)

9781479966820

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2015 Institute of Electrical and Electronics Engineers (IEEE), All rights reserved.

Share

 
COinS