Diagnosis of Paratuberculosis in Histopathological Images Based on Explainable Artificial Intelligence and Deep Learning

Abstract

Artificial intelligence holds great promise in medical imaging, especially histopathological imaging. However, artificial intelligence algorithms cannot fully explain the thought processes during decision-making. This situation has brought the problem of explainability, i.e., the black box problem, of artificial intelligence applications to the agenda: an algorithm simply responds without stating the reasons for the given images. To overcome the problem and improve the explainability, explainable artificial intelligence (XAI) has come to the fore, and piqued the interest of many researchers. Against this backdrop, this study examines a new and original dataset using the deep learning algorithm, and visualizes the output with gradient-weighted class activation mapping (Grad-CAM), one of the XAI applications. Afterwards, a detailed questionnaire survey was conducted with the pathologists on these images. Both the decision-making processes and the explanations were verified, and the accuracy of the output was tested. The research results greatly help pathologists in the diagnosis of paratuberculosis.

Increasing Trust in AI Using Explainable Artificial Intelligence for Histopathology – An Overview

“…From this review the conclusion is that each method offers some insight on the decision. For example Grad-CAM from [3] offers the gradients of any target concept flowing into the final convolutional layer of the CNN to produce a coarse localization map that highlights the important regions in the image for predicting the concept. From the methods described in this review we could infer that there is a need for multiple parameters to be understood by a histopathologist.…”

“…Another method will be the heatmap visualization. The Grad-CAM method was also used by [3]. A paper that refers to cancer diagnostics [4] emphasizes that the acceptance of deep learning is conditioned by the ability to understand what the algorithms are doing.…”

Improving Tuberculosis Diagnosis using Explainable Artificial Intelligence in Medical Imaging

“…They have examined a new dataset using deep learning algorithms and visualized outputs with Grad-CAM, demonstrating the application of XAI in diagnosing paratuberculosis from histopathological images [12].…”