Memory transformers for full context and high-resolution 3D Medical Segmentation

TL;DR

This paper introduces the FINE transformer, which uses memory tokens to enable full-range attention in high-resolution 3D medical image segmentation, outperforming existing CNN and transformer models.

Contribution

The FINE transformer is a novel architecture that scales full attention to high-resolution 3D images using memory tokens at two levels, improving segmentation performance.

Findings

FINE outperforms state-of-the-art CNN and transformer baselines.

Memory tokens enable full-range interactions in high-resolution 3D images.

FINE achieves superior results on the BCV dataset.

Abstract

Transformer models achieve state-of-the-art results for image segmentation. However, achieving long-range attention, necessary to capture global context, with high-resolution 3D images is a fundamental challenge. This paper introduces the Full resolutIoN mEmory (FINE) transformer to overcome this issue. The core idea behind FINE is to learn memory tokens to indirectly model full range interactions while scaling well in both memory and computational costs. FINE introduces memory tokens at two levels: the first one allows full interaction between voxels within local image regions (patches), the second one allows full interactions between all regions of the 3D volume. Combined, they allow full attention over high resolution images, e.g. 512 x 512 x 256 voxels and above. Experiments on the BCV image segmentation dataset shows better performances than state-of-the-art CNN and transformer baselines, highlighting the superiority of our full attention mechanism compared to recent transformer baselines, e.g. CoTr, and nnFormer.