Enhancing bone radiology images classification through appropriate preprocessing: a deep learning and explainable artificial intelligence approach.

ABSTRACT:

Background

Medical image classification has been an important application for deep learning techniques for over a decade, and since the emergence of explainable artificial intelligence (XAI), researchers have started using XAI to validate the results produced by these black box models. In the research field, it has become clear that accuracy and efficiency are not the only crucial factors for developing medical deep learning models; the authenticity of results and the accountability of the model and its creator also matter greatly. The objective of this study is to emphasize the importance of authenticity of the results and the accountability of deep learning models used for medical purposes, through proposing targeted preprocessing method for medical dataset processed by deep lear

SUBMITTER: Wu Y 

PROVIDER: S-EPMC11948367 | biostudies-literature | 2025 Mar

REPOSITORIES: biostudies-literature