Towards an interpretable model for automatic classification of endoscopy images

Deep learning strategies have become the mainstream for computer-assisted diagnosis tools development since they outperform other machine learning techniques. However, these systems can not reach their full potential since the lack of understanding of their operation and questionable generalizability provokes mistrust from the users, limiting their application. In this paper, we generate a Convolutional Neural Network (CNN) using a genetic algorithm for hyperparameter optimization. Our CNN has state-of-the-art classification performance, delivering higher evaluation metrics than other recent papers that use AI models to classify images from the same dataset. We provide visual explanations of the classifications made by our model implementing Grad-CAM and analyze the behavior of our model on misclassifications using this technique.

Citation

García-Aguirre, R., Torres Treviño, L., Navarro-López, E.M. and González-González, J.A. (2022) Towards an interpretable model for automatic classification of endoscopy images. Lecture Notes in Artificial Intelligence, 13612, pp. 297-307. DOI: 10.1007/978-3-031-19493-1_24

Type

Conference contribution

Language

en

Description

This is an accepted manuscript of a paper published by Springer in Lecture Notes in Artificial Intelligence on 23 October 2022. Conference paper presented at the 21st Mexican International Conference on Artificial Intelligence, 24th-29th October 2022. For re-use please see the publisher's terms and conditions.

Series/Report no.

Lecture Notes in Artificial Intelligence
Lecture Notes in Computer Science

ISSN

0302-9743

EISSN

1611-3349

ISBN

9783031194948

Collections

Faculty of Science and Engineering