Locally-Supervised Global Image Restoration

2025-11-03

AI Summary

This work addresses image reconstruction under fixed, deterministic incomplete sampling—such as super-resolution and inpainting—by proposing a self-supervised learning framework that requires no fully labeled data. Unlike conventional methods relying on stochastic sampling or fully observed ground-truth labels, our approach introduces, for the first time, multi-invariance priors (e.g., translation, rotation) and theoretically proves that global reconstruction is achievable from local observations alone, attaining performance on par with fully supervised methods. The framework trains deep neural networks via a localized sampling strategy that explicitly leverages geometric invariances inherent in natural image distributions. Evaluated on photoacoustic microscopy super-resolution, our method matches or surpasses fully supervised baselines while drastically reducing dependence on complete ground-truth annotations.

Technology Category

• Computer Vision: Learning & Optimization for CV
• 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
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Security and Privacy: Data transparency and provenance

Abstract

We address the problem of image reconstruction from incomplete measurements, encompassing both upsampling and inpainting, within a learning-based framework. Conventional supervised approaches require fully sampled ground truth data, while self-supervised methods allow incomplete ground truth but typically rely on random sampling that, in expectation, covers the entire image. In contrast, we consider fixed, deterministic sampling patterns with inherently incomplete coverage, even in expectation. To overcome this limitation, we exploit multiple invariances of the underlying image distribution, which theoretically allows us to achieve the same reconstruction performance as fully supervised approaches. We validate our method on optical-resolution image upsampling in photoacoustic microscopy (PAM), demonstrating competitive or superior results while requiring substantially less ground truth data.

Problem

• Image reconstruction from incomplete measurements with deterministic sampling
• Overcoming limited coverage using multiple invariances of image distribution
• Validating method on optical-resolution upsampling in photoacoustic microscopy

Innovation

• Locally-supervised global image restoration framework
• Exploiting multiple invariances of image distribution
• Fixed deterministic sampling with incomplete coverage