Research on improving pedestrian detection algorithm based on YOLOv5

Xiaogang Lin; Anjun Song

doi:10.1117/12.2682285

26 May 2023 Research on improving pedestrian detection algorithm based on YOLOv5

Xiaogang Lin, Anjun Song

Proceedings Volume 12700, International Conference on Electronic Information Engineering and Data Processing (EIEDP 2023); 1270024 (2023) https://doi.org/10.1117/12.2682285
Event: International Conference on Electronic Information Engineering and Data Processing (EIEDP 2023), 2023, Nanchang, China

Abstract

We proposed a pedestrian detection algorithm combining YOLOv5 with convolution and channel attention mechanism. First, we use our own pedestrian dataset to train the YOLOv5 detection model. Then, three attention mechanisms, SE, CBAMC3, and CoordAtt, are used to enhance the detection performance. The experiment demonstrated that the precision of both SE and CoordAtt decreased, the recall of SE also decreased, while the accuracy of CBAMC3 was improved and the mAP changed little, thus CBAMC3 became the best model for pedestrian detection. Research indicates that adding convolution block attention modules can increase the precision of detecting small pedestrian targets.

Citation Download Citation

Xiaogang Lin and Anjun Song "Research on improving pedestrian detection algorithm based on YOLOv5", Proc. SPIE 12700, International Conference on Electronic Information Engineering and Data Processing (EIEDP 2023), 1270024 (26 May 2023); https://doi.org/10.1117/12.2682285

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available