Pressure injury ulcer stage classification based on convolution neural network image recognition

Ying-Jui Huang; Wei-Cheng Hung; I-Hsu Hsu; Chao-Teng Su; Fu-Li Hsiao

doi:10.1117/12.3036149

31 December 2024 Pressure injury ulcer stage classification based on convolution neural network image recognition

Ying-Jui Huang, Wei-Cheng Hung, I-Hsu Hsu, Chao-Teng Su, Fu-Li Hsiao

Author Affiliations +

Proceedings Volume 13487, Optics and Photonics International Congress 2024; 1348709 (2024) https://doi.org/10.1117/12.3036149
Event: Optics and Photonics International Congress, 2024, Yokohama, Japan

Abstract

In this study, we employed a dataset of 227 photographs of pressure injuries of various grades to train a range of pretrained convolutional neural network architectures for wound grade classification. These architectures included six types: GoogLeNet, ResNet18, ResNet50, ResNet101, VGG-19, and VGG-16. We fine-tuned parameters such as maximum batch size and maximum epochs to optimize resolution. Concurrently, we utilized the results from Gradient Class Activation Mapping to analyze the reasonableness of each optimized result. Our findings indicate that among the assessed architectures, GoogLeNet is the most suitable for clinical application in a comprehensive evaluation.

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

Citation Download Citation

Ying-Jui Huang, Wei-Cheng Hung, I-Hsu Hsu, Chao-Teng Su, and Fu-Li Hsiao "Pressure injury ulcer stage classification based on convolution neural network image recognition", Proc. SPIE 13487, Optics and Photonics International Congress 2024, 1348709 (31 December 2024); https://doi.org/10.1117/12.3036149

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