A study on the detection of breast lumps based on attentional mechanisms

Lanfeng Zhou; Zhikun Chen

doi:10.1117/12.2687392

28 August 2023 A study on the detection of breast lumps based on attentional mechanisms

Lanfeng Zhou, Zhikun Chen

Proceedings Volume 12724, Second International Conference on Biomedical and Intelligent Systems (IC-BIS 2023); 127241N (2023) https://doi.org/10.1117/12.2687392
Event: Second International Conference on Biomedical and Intelligent Systems (IC-BIS2023), 2023, Xiamen, China

Abstract

Traditional networks cannot focus on which features are important when extracting features and use the same weighting for all features, so this paper investigates breast lump detection based on attention mechanisms. By adding three attention mechanisms, SE, ECA and CBAM, respectively, to the backbone network of YOLOv5s, the network's ability to extract breast lump features is enhanced for breast lump recognition. The experimental results show that adding SE or ECA or CBAM attention mechanisms alone to the YOLOv5s network, they are higher in accuracy, recall and mAP@0.5 than the original YOLOv5s network model, and the model with the addition of CBAM attention mechanism improves the accuracy, recall and mAP@0.5 by 3.1%, 2.6% and 3.1% respectively, which is better than the models with the addition of SE and ECA attention mechanisms. The proposed attention mechanism-based breast lump detection model can effectively detect breast lumps and provide help for early screening.

Citation Download Citation

Lanfeng Zhou and Zhikun Chen "A study on the detection of breast lumps based on attentional mechanisms", Proc. SPIE 12724, Second International Conference on Biomedical and Intelligent Systems (IC-BIS 2023), 127241N (28 August 2023); https://doi.org/10.1117/12.2687392

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