Reliable and trustworthy deep learning algorithm design for oral cancer image analysis

Bofan Song; Dharma KC; Shaobai Li; Chicheng Zhang; Rongguang Liang

doi:10.1117/12.3007296

13 March 2024 Reliable and trustworthy deep learning algorithm design for oral cancer image analysis

Bofan Song, Dharma KC, Shaobai Li, Chicheng Zhang, Rongguang Liang

Proceedings Volume PC12833, Design and Quality for Biomedical Technologies XVII; PC1283302 (2024) https://doi.org/10.1117/12.3007296
Event: SPIE BiOS, 2024, San Francisco, California, United States

Abstract

Deep learning is a powerful tool for image analysis and medical applications. However, due to their intricate black-box nature, comprehending deep learning model predictions are often challenging. Oral cancer is globally prevalent, necessitating reliable AI algorithms for screening, especially for low-income regions. Interpretability is crucial for reliable AI. Visual explanation, generating attention maps highlighting decision-influencing regions, aids interpretability and also helps guide AI focus. Elevating AI reliability involves assessing decision confidence as well. Quantifying model output certainty helps identify uncertain cases, which need additional examination. Dataset quality is also pivotal for reliable AI development. Methods to evaluate and enhance the data and label/annotation quality will also be essential.

Conference Presentation

Citation Download Citation

Bofan Song, Dharma KC, Shaobai Li, Chicheng Zhang, and Rongguang Liang "Reliable and trustworthy deep learning algorithm design for oral cancer image analysis", Proc. SPIE PC12833, Design and Quality for Biomedical Technologies XVII, PC1283302 (13 March 2024); https://doi.org/10.1117/12.3007296

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