Occluded Pedestrian Detection with Visible IoU and Box Sign Predictor

TL;DR

This paper introduces visible IoU and a box sign predictor to improve occluded pedestrian detection, achieving state-of-the-art results by explicitly considering occlusion and predicting movement directions.

Contribution

It proposes visible IoU for better sample selection and a box sign predictor for improved localization, advancing occluded pedestrian detection methods.

Findings

Achieved state-of-the-art performance on CityPersons benchmark.

Visible IoU improves training sample quality for occluded objects.

Sign predictor enhances localization accuracy through direction estimation.

Abstract

Training a robust classifier and an accurate box regressor are difficult for occluded pedestrian detection. Traditionally adopted Intersection over Union (IoU) measurement does not consider the occluded region of the object and leads to improper training samples. To address such issue, a modification called visible IoU is proposed in this paper to explicitly incorporate the visible ratio in selecting samples. Then a newly designed box sign predictor is placed in parallel with box regressor to separately predict the moving direction of training samples. It leads to higher localization accuracy by introducing sign prediction loss during training and sign refining in testing. Following these novelties, we obtain state-of-the-art performance on CityPersons benchmark for occluded pedestrian detection.