Fast and Accurate Gigapixel Pathological Image Classification with Hierarchical Distillation Multi-Instance Learning

TL;DR

HDMIL is a hierarchical distillation framework that accelerates gigapixel pathological image classification by efficiently identifying relevant regions, reducing inference time, and improving accuracy through a novel combination of high- and low-resolution analysis.

Contribution

This paper introduces HDMIL, a novel hierarchical distillation multi-instance learning framework with a dual-network approach and a Chebyshev-polynomials-based classifier for faster, more accurate pathological image classification.

Findings

HDMIL achieves 3.13% higher AUC on Camelyon16.

Inference time is reduced by 28.6%.

Outperforms previous state-of-the-art methods.

Abstract

Although multi-instance learning (MIL) has succeeded in pathological image classification, it faces the challenge of high inference costs due to processing numerous patches from gigapixel whole slide images (WSIs). To address this, we propose HDMIL, a hierarchical distillation multi-instance learning framework that achieves fast and accurate classification by eliminating irrelevant patches. HDMIL consists of two key components: the dynamic multi-instance network (DMIN) and the lightweight instance pre-screening network (LIPN). DMIN operates on high-resolution WSIs, while LIPN operates on the corresponding low-resolution counterparts. During training, DMIN are trained for WSI classification while generating attention-score-based masks that indicate irrelevant patches. These masks then guide the training of LIPN to predict the relevance of each low-resolution patch. During testing, LIPN first determines the useful regions within low-resolution WSIs, which indirectly enables us to eliminate irrelevant regions in high-resolution WSIs, thereby reducing inference time without causing performance degradation. In addition, we further design the first Chebyshev-polynomials-based Kolmogorov-Arnold classifier in computational pathology, which enhances the performance of HDMIL through learnable activation layers. Extensive experiments on three public datasets demonstrate that HDMIL outperforms previous state-of-the-art methods, e.g., achieving improvements of 3.13% in AUC while reducing inference time by 28.6% on the Camelyon16 dataset.