Improved YOLO framework for cell detection in bronchoalveolar lavage fluid

Xin Zhang; Rui Li; Chao Wang; Yujuan Jia

doi:10.1117/12.3010582

1 December 2023 Improved YOLO framework for cell detection in bronchoalveolar lavage fluid

Xin Zhang, Rui Li, Chao Wang, Yujuan Jia

Proceedings Volume 12940, Third International Conference on Control and Intelligent Robotics (ICCIR 2023); 129403B (2023) https://doi.org/10.1117/12.3010582
Event: Third International Conference on Control and Intelligent Robotics (ICCIR 2023), 2023, Sipsongpanna, China

Abstract

In response to the limited amount of cell detection data and the phenomenon of overlapping and adhesion in the existing bronchoalveolar lavage fluid (BALF), a YOLOv5s-based target detection algorithm (YOLOv5s-GAN) was proposed to optimize the model structure and loss function and effectively address the overlapping and adhesion of BALF cells. The algorithm enhances the accuracy of cell detection in BALF by using a lightweight convolution GSconv module and a normalization-based attention mechanism (NAS) that does not require additional calculations or parameters. Additionally, the algorithm improves detection accuracy by using D-IoU to mitigate false detections due to overlapping cells. Experimental results showed that the algorithm achieved an average accuracy of 88.9% in detecting four types of BALF cells, which demonstrated its high practicality for cell detection in BALF.

(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Xin Zhang, Rui Li, Chao Wang, and Yujuan Jia "Improved YOLO framework for cell detection in bronchoalveolar lavage fluid", Proc. SPIE 12940, Third International Conference on Control and Intelligent Robotics (ICCIR 2023), 129403B (1 December 2023); https://doi.org/10.1117/12.3010582

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