Mamba Goes HoME: Hierarchical Soft Mixture-of-Experts for 3D Medical Image Segmentation

TL;DR

This paper introduces HoME, a hierarchical mixture-of-experts model that improves 3D medical image segmentation by efficiently capturing long-range dependencies and handling diverse data modalities.

Contribution

It proposes a novel two-level token-routing architecture built on the Mamba SSM backbone, enhancing segmentation accuracy and generalizability across multiple 3D medical imaging modalities.

Findings

Surpasses state-of-the-art segmentation performance on multiple datasets

Efficiently models long-range dependencies in 3D medical images

Demonstrates robustness across diverse data qualities and modalities

Abstract

In recent years, artificial intelligence has significantly advanced medical image segmentation. Nonetheless, challenges remain, including efficient 3D medical image processing across diverse modalities and handling data variability. In this work, we introduce Hierarchical Soft Mixture-of-Experts (HoME), a two-level token-routing layer for efficient long-context modeling, specifically designed for 3D medical image segmentation. Built on the Mamba Selective State Space Model (SSM) backbone, HoME enhances sequential modeling through adaptive expert routing. In the first level, a Soft Mixture-of-Experts (SMoE) layer partitions input sequences into local groups, routing tokens to specialized per-group experts for localized feature extraction. The second level aggregates these outputs through a global SMoE layer, enabling cross-group information fusion and global context refinement. This hierarchical design, combining local expert routing with global expert refinement, enhances generalizability and segmentation performance, surpassing state-of-the-art results across datasets from the three most widely used 3D medical imaging modalities and varying data qualities. The code is publicly available at https://github.com/gmum/MambaHoME.