SELMA3D challenge: Self-supervised learning for 3D light-sheet microscopy image segmentation

2025-01-07

AI Summary

Domain shift severely limits generalizability in 3D light-sheet microscopy (LSM) image segmentation of large-scale biological tissues. Method: We introduce the first MICCAI challenge dataset tailored for self-supervised learning, comprising 35 ultra-large unlabeled 3D volumes (>1000³ voxels) of mouse and human brain tissue and 315 finely annotated subvolumes. We pioneer the integration of contrastive learning and masked autoencoding for 3D LSM segmentation, proposing a novel cross-sample, cross-structure (vasculature-/point-like) representation learning paradigm. Contribution/Results: This establishes the first biomedical 3D self-supervised segmentation benchmark. Participating methods include ViT- and U-Net-based architectures and voxel-level reconstruction tasks. On unseen test data, our approach achieves an average Dice score improvement of 12.6% over fully supervised baselines—demonstrating that large-scale unlabeled 3D biomedical imagery significantly enhances robustness and generalizability of segmentation models.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies

Abstract

Recent innovations in light sheet microscopy, paired with developments in tissue clearing techniques, enable the 3D imaging of large mammalian tissues with cellular resolution. Combined with the progress in large-scale data analysis, driven by deep learning, these innovations empower researchers to rapidly investigate the morphological and functional properties of diverse biological samples. Segmentation, a crucial preliminary step in the analysis process, can be automated using domain-specific deep learning models with expert-level performance. However, these models exhibit high sensitivity to domain shifts, leading to a significant drop in accuracy when applied to data outside their training distribution. To address this limitation, and inspired by the recent success of self-supervised learning in training generalizable models, we organized the SELMA3D Challenge during the MICCAI 2024 conference. SELMA3D provides a vast collection of light-sheet images from cleared mice and human brains, comprising 35 large 3D images-each with over 1000^3 voxels-and 315 annotated small patches for finetuning, preliminary testing and final testing. The dataset encompasses diverse biological structures, including vessel-like and spot-like structures. Five teams participated in all phases of the challenge, and their proposed methods are reviewed in this paper. Quantitative and qualitative results from most participating teams demonstrate that self-supervised learning on large datasets improves segmentation model performance and generalization. We will continue to support and extend SELMA3D as an inaugural MICCAI challenge focused on self-supervised learning for 3D microscopy image segmentation.

Problem

• 3D microscopy image segmentation
• accuracy degradation
• adaptability

Innovation

• Self-supervised Learning
• 3D Microscopy Image Segmentation
• SELMA3D Challenge