DA-Net: Pedestrian Detection Using Dense Connected Block and Attention Modules

Ruihong Yin,Wei Zhao,Feng Jiang,Rufei Zhang

doi:10.1109/access.2020.3018306

Ruihong Yin, Wei Zhao + Show 2 more

Open Access

https://doi.org/10.1109/access.2020.3018306

Copy DOI

Abstract

Pedestrian detection plays an important role in some areas such as autonomous driving, but due to heavy occlusion and various scales, it is still challenging. In this article, we propose an improved pedestrian detection method called DA-Net based on the two-stage detector Feature Pyramid Network (FPN). DA-Net adds Dense Connected Block (DCB), a combination of channel-wise attention module (CWAM) and global attention module (GAM) to the network. FPN can produce features with various scales and semantic information, which is good for the detection of pedestrians on various scales. Due to many small-scale targets in pedestrian detection, we only regard the low layers with enough details of targets in FPN as prediction layers. After several DCBs to deepen the network, prediction layers in our network can encode richer semantic information of targets, which can make the location of a target more precisely. In order to highlight visible parts of occluded pedestrians and ignore occluded parts, CWAM weights each channel of features with different importance. GAM aggregates global information and long-range dependencies for small-scale and occluded targets. Thus, the combination of CWAM and GAM is not only beneficial for coping with occlusion problem in pedestrian detection, but also for gaining environmental information for small-scale targets. Evaluation results on CUHK and CityPersons datasets show that our proposed method achieves improved performance with log-average miss rate reduction of 9.6% on the CUHK dataset and 6.1% on the Heavy subset of CityPersons dataset compared with FPN.

Highlights

Object detection is one of the essential research fields in computer vision, whose task is to find all objects in an image
2) INFLUENCE OF THE COMBINATION OF channel-wise attention module (CWAM) AND global attention module (GAM) After adding Dense Connected Block (DCB) for richer semantic information, we add a combination of CWAM and GAM after {C2new, C3new, C4new, C5new} for getting channel-wise and environmental information
The results show that our DCB and the combination of CWAM and GAM are both valid for pedestrian detection

Summary

INTRODUCTION

Object detection is one of the essential research fields in computer vision, whose task is to find all objects in an image. Two-stage algorithms extract the region proposals first and locate the targets, so detection accuracy is high at the expense of time. FPN[12] designs a top-down architecture with lateral connections for building feature maps at multiple scales It can compensate for missing information in the down-sampling process of CNNs. It can compensate for missing information in the down-sampling process of CNNs Another challenging problem in pedestrian detection is that there is a large variation of pedestrian scales and the number of small-scale targets is especially high. We propose a network (DA-Net) using Dense Connected Block and attention modules for pedestrian detection. We adopt Feature Pyramid Network (FPN) with ResNet50[25] to face with various scales in pedestrian detection and only use the low layers for prediction.

THE PROPOSED METHOD

DENSE CONNECTED BLOCK

RESULTS

CONCLUSION