SC-MIL: Sparsely Coded Multiple Instance Learning for Whole Slide Image Classification

TL;DR

This paper introduces SC-MIL, a novel sparsely coded multiple instance learning method that enhances whole slide image classification by capturing instance similarities and suppressing irrelevant features through sparse dictionary learning and deep unrolling.

Contribution

The paper proposes a plug-and-play SC module using sparse dictionary learning and deep unrolling to improve MIL performance in WSI classification, addressing feature embedding and instance correlation modeling simultaneously.

Findings

SC-MIL significantly improves state-of-the-art MIL methods.

The sparse coding approach effectively suppresses irrelevant instances.

Experimental results show substantial performance boosts across multiple datasets.

Abstract

Multiple Instance Learning (MIL) has been widely used in weakly supervised whole slide image (WSI) classification. Typical MIL methods include a feature embedding part, which embeds the instances into features via a pre-trained feature extractor, and an MIL aggregator that combines instance embeddings into predictions. Most efforts have typically focused on improving these parts. This involves refining the feature embeddings through self-supervised pre-training as well as modeling the correlations between instances separately. In this paper, we proposed a sparsely coding MIL (SC-MIL) method that addresses those two aspects at the same time by leveraging sparse dictionary learning. The sparse dictionary learning captures the similarities of instances by expressing them as sparse linear combinations of atoms in an over-complete dictionary. In addition, imposing sparsity improves instance feature embeddings by suppressing irrelevant instances while retaining the most relevant ones. To make the conventional sparse coding algorithm compatible with deep learning, we unrolled it into a sparsely coded module leveraging deep unrolling. The proposed SC module can be incorporated into any existing MIL framework in a plug-and-play manner with an acceptable computational cost. The experimental results on multiple datasets demonstrated that the proposed SC module could substantially boost the performance of state-of-the-art MIL methods. The codes are available at \href{https://github.com/sotiraslab/SCMIL.git}{https://github.com/sotiraslab/SCMIL.git}.