3D MRI Image Pretraining via Controllable 2D Slice Navigation Task

TL;DR

This paper introduces a novel self-supervised pretraining method for MRI that transforms 3D volumes into controllable 2D slice sequences, enhancing the learning of spatial and anatomical features.

Contribution

It proposes a new pretraining approach using controllable 2D slice navigation and action-conditioned objectives, differing from traditional static volume methods.

Findings

Pretraining with controllable slice navigation improves downstream task performance.

The method outperforms static-volume baselines and other pretraining variants.

Controllable MRI slice navigation offers a complementary interface for representation learning.

Abstract

Self-supervised pretraining has become the mainstream approach for learning MRI representations from unlabeled scans. However, most existing objectives still treat each scan primarily as static aggregations of slices, patches or volumes. We ask whether there exists an intrinsic form of self-supervision signal that is different from reconstructing the masked patches, through transforming the 3D volumes into controllable 2D rendered sequences: by rendering slices at continuous positions, orientations, and scales, a 3D volume can be converted into dense video-action sequences whose controls are the action trajectories. We study this formulation with an action-conditioned pretraining objective, where a tokenizer encodes slice observations and a latent dynamics model predicts the evolution of latent features. Across representative anatomical and spatial downstream tasks, the proposed pretraining is evaluated against standard static-volume baselines, tokenizer-only pretraining, and dynamics variants without aligned actions. These results suggest that controllable MRI slice navigation provides a useful complementary pretraining interface for learning anatomical and spatial representations from large unlabeled MRI collections.