Object detection algorithm based on improved YOLOv5

Linfeng Cai; Bin Tang; Yifei Xu; SiYue Lei; Yulong He; Jinfu Zhang; Zourong Long

doi:10.1117/12.2656595

20 October 2022 Object detection algorithm based on improved YOLOv5

Linfeng Cai, Bin Tang, Yifei Xu, SiYue Lei, Yulong He, Jinfu Zhang, Zourong Long

Proceedings Volume 12451, 5th International Conference on Computer Information Science and Application Technology (CISAT 2022); 124514H (2022) https://doi.org/10.1117/12.2656595
Event: 5th International Conference on Computer Information Science and Application Technology (CISAT 2022), 2022, Chongqing, China

Abstract

One of the research directions of target detection-based computer vision, in which small target detection is the key and difficult research direction in target detection. Traditional target detection algorithms include Faster RCNN, YOLO, SSD, etc., and there is a problem that indicators such as detection accuracy, false detection rate, and missed detection rate are not ideal for small target detection tasks. In order to improve the above problems, this paper proposes an improved target detection algorithm based on YOLOv5. First, the CBAM attention mechanism is introduced in the Backbone part to strengthen the important feature channels; then a detection layer is added to the network according to the characteristics of the data set to strengthen the extraction ability. Experiments show that the improved YOLOv5s_CS algorithm has a mAP value of 75.1% on the test set, which is 3.9% higher than the original YOLOv5s network.

Citation Download Citation

Linfeng Cai, Bin Tang, Yifei Xu, SiYue Lei, Yulong He, Jinfu Zhang, and Zourong Long "Object detection algorithm based on improved YOLOv5", Proc. SPIE 12451, 5th International Conference on Computer Information Science and Application Technology (CISAT 2022), 124514H (20 October 2022); https://doi.org/10.1117/12.2656595

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