Effective-LDAM: An Effective Loss Function To Mitigate Data Imbalance for Robust Chest X-Ray Disease Classification
Sree Rama Vamsidhar S, Bhargava Satya, Rama Krishna Gorthi
TL;DR
The paper addresses the challenge of data imbalance in chest X-ray disease classification, particularly for COVID-19 detection, by introducing Effective-LDAM (E-LDAM), a margin-based loss that leverages the effective number of samples to tailor class margins. E-LDAM modifies the LDAM formulation by computing class-specific margins with $\Delta_j=\frac{C}{E_{n_j}^{1/r}}$, where the effective number of samples is $E_n=\frac{1-\beta^n}{1-\beta}$, thereby enhancing minority-class discrimination. The authors implement E-LDAM within a Heat Guided Convolutional Neural Network (HG-CNN) architecture and evaluate on the COVIDx dataset, achieving a COVID-19 recall of $97.81\%$ and an overall accuracy around the mid-90s, outperforming CE, CB-CE, and LDAM. This algorithmic approach offers a robust alternative to data augmentation for addressing imbalanced medical imaging data and can improve diagnostic reliability for data-scarce infections. The work demonstrates the practical impact of margin-based loss design in medical imaging and suggests avenues for applying effective-number adjustments to other imbalanced domains.
Abstract
Deep Learning (DL) approaches have gained prominence in medical imaging for disease diagnosis. Chest X-ray (CXR) classification has emerged as an effective method for detecting various diseases. Among these methodologies, Chest X-ray (CXR) classification has proven to be an effective approach for detecting and analyzing various diseases. However, the reliable performance of DL classification algorithms is dependent upon access to large and balanced datasets, which pose challenges in medical imaging due to the impracticality of acquiring sufficient data for every disease category. To tackle this problem, we propose an algorithmic-centric approach called Effective-Label Distribution Aware Margin (E-LDAM), which modifies the margin of the widely adopted Label Distribution Aware Margin (LDAM) loss function using an effective number of samples in each class. Experimental evaluations on the COVIDx CXR dataset focus on Normal, Pneumonia, and COVID-19 classification. The experimental results demonstrate the effectiveness of the proposed E-LDAM approach, achieving a remarkable recall score of 97.81% for the minority class (COVID-19) in CXR image prediction. Furthermore, the overall accuracy of the three-class classification task attains an impressive level of 95.26%.