Chest X-rays Classification: A Multi-Label and Fine-Grained Problem

TL;DR

This paper introduces a novel multi-label softmax loss function and a fine-grained classification architecture to improve chest X-ray image classification, addressing challenges like multiple labels, class imbalance, and similar visual features.

Contribution

The paper proposes a new error function, MSML, and a fine-grained network design that enhance multi-label classification performance on chest X-ray datasets.

Findings

Consistent AUC-ROC improvements across various network backbones.

MSML outperforms traditional loss functions in multi-label, imbalanced data scenarios.

Enhanced detection of multiple lung pathologies in chest X-ray images.

Abstract

The widely used ChestX-ray14 dataset addresses an important medical image classification problem and has the following caveats: 1) many lung pathologies are visually similar, 2) a variant of diseases including lung cancer, tuberculosis, and pneumonia are present in a single scan, i.e. multiple labels and 3) The incidence of healthy images is much larger than diseased samples, creating imbalanced data. These properties are common in medical domain. Existing literature uses stateof- the-art DensetNet/Resnet models being transfer learned where output neurons of the networks are trained for individual diseases to cater for multiple diseases labels in each image. However, most of them don't consider relationship between multiple classes. In this work we have proposed a novel error function, Multi-label Softmax Loss (MSML), to specifically address the properties of multiple labels and imbalanced data. Moreover, we have designed deep network architecture based on fine-grained classification concept that incorporates MSML. We have evaluated our proposed method on various network backbones and showed consistent performance improvements of AUC-ROC scores on the ChestX-ray14 dataset. The proposed error function provides a new method to gain improved performance across wider medical datasets.