Towards Interpretable Classification of Leukocytes based on Deep Learning

TL;DR

This paper explores the calibration of confidence estimates and visual explanation methods for deep learning-based leukocyte classification, aiming to improve interpretability and clinical integration of label-free cytological imaging.

Contribution

It introduces methods for confidence calibration and compares visual explanation techniques to enhance interpretability in leukocyte classification.

Findings

Effective confidence calibration improves trust in model predictions.

Visual explanation methods help interpret neural network decisions.

Identified general detection patterns in neural networks for blood cell analysis.

Abstract

Label-free approaches are attractive in cytological imaging due to their flexibility and cost efficiency. They are supported by machine learning methods, which, despite the lack of labeling and the associated lower contrast, can classify cells with high accuracy where the human observer has little chance to discriminate cells. In order to better integrate these workflows into the clinical decision making process, this work investigates the calibration of confidence estimation for the automated classification of leukocytes. In addition, different visual explanation approaches are compared, which should bring machine decision making closer to professional healthcare applications. Furthermore, we were able to identify general detection patterns in neural networks and demonstrate the utility of the presented approaches in different scenarios of blood cell analysis.