Improved small object detection algorithm based on YOLOv5

Shenshen Sun; Xirui Wang; Xue Bao

doi:10.1117/12.3026537

27 March 2024 Improved small object detection algorithm based on YOLOv5

Shenshen Sun, Xirui Wang, Xue Bao

Proceedings Volume 13105, International Conference on Computer Graphics, Artificial Intelligence, and Data Processing (ICCAID 2023); 131054J (2024) https://doi.org/10.1117/12.3026537
Event: 3rd International Conference on Computer Graphics, Artificial Intelligence, and Data Processing (ICCAID 2023), 2023, Qingdao, China

Abstract

In order to address the issues of significant target scale variations and insufficient feature information in road scenarios, an enhanced model named SC-YOLO based on the YOLOv5s algorithm is proposed. The model utilizes the Swin Transformer Block to expand the receptive field and strengthen the feature extraction capability. Leveraging the ConvNeXt Block, the model promotes feature information fusion to mitigate missed detections and false positives. Experimental evaluation on the KITTI traffic object dataset demonstrates that the improved model yields notable enhancements in terms of mAP@50%, accuracy, and recall rate, with improvements of 6.2%, 8.5%, and 4.8%, respectively, compared to the original algorithm. These results affirm the effectiveness of the improved model in enhancing the detection performance of small targets in complex road scenarios.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Shenshen Sun, Xirui Wang, and Xue Bao "Improved small object detection algorithm based on YOLOv5", Proc. SPIE 13105, International Conference on Computer Graphics, Artificial Intelligence, and Data Processing (ICCAID 2023), 131054J (27 March 2024); https://doi.org/10.1117/12.3026537

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