SkelEM: Training-Signal Decoupling of Skeleton and Diffusion for Self-supervised Axial Super-Resolution in Volume Microscopy

2026-06-29

AI Summary

This work addresses the anisotropic resolution in volumetric microscopy caused by limited axial sampling, a challenge that existing self-supervised super-resolution methods struggle to overcome while balancing smoothness, structural fidelity, and inference efficiency. The authors propose SkelEM, a novel framework that decouples skeleton extraction from diffusion-based refinement at the training signal level: a frozen topological network generates low-frequency structural priors, while a diffusion model focuses exclusively on enhancing high-frequency details. Realistic residual priors are extracted via cycle consistency on sparse slices to guide artifact removal. Coupled with truncated reverse diffusion (≤5 steps), the method enables efficient, high-fidelity reconstruction. SkelEM achieves state-of-the-art fidelity–perception trade-offs across multiple benchmarks, demonstrates superior performance in downstream membrane segmentation, exhibits strong zero-shot cross-modality generalization, and introduces the BRAVE-ASR benchmark for cross-instrument evaluation.

Abstract

Volume microscopy, including electron and light microscopy, suffers from severe anisotropic resolution due to physical axial sectioning. Existing self-supervised axial super-resolution (ASR) methods face a trilemma bounded by overly smoothed regression textures, structural hallucinations of pure diffusion models, and prohibitive inference latency. In this paper, we propose Skeleton-refinE Microscopy (SkelEM), a self-supervised framework that decouples ASR at the training-signal level: a frozen topological network and a diffusion refiner are optimized by disjoint objectives, separating low-frequency topology formulation from high-frequency detail enhancement. Building on this deterministic skeleton, we exploit a unified cycle-consistent mechanism on input sparse slices to simultaneously extract a real-domain residual prior and bidirectionally align the diffusion refiner, washing away cross-plane artifacts without synthetic bias. By truncating the reverse diffusion process with this physical prior, SkelEM achieves high-fidelity detail restoration in merely $\le 5$ steps. To rigorously assess cross-instrument generalization, we further introduce BRAVE-ASR, a new benchmark of co-aligned anisotropic and isotropic volumes acquired on a Plasma-FIB instrument. Across public benchmarks, SkelEM achieves the most favorable balance across the fidelity-perception trade-off among self-supervised methods, with state-of-the-art downstream membrane segmentation performance and robust zero-shot generalization across distinct modalities.

Problem

• axial super-resolution
• anisotropic resolution
• volume microscopy
• self-supervised learning
• structural hallucination

Innovation

• training-signal decoupling
• self-supervised axial super-resolution
• diffusion refiner
• cycle-consistent mechanism
• topological skeleton