CVMar 16, 2019

Robust Super-Resolution GAN, with Manifold-based and Perception Loss

arXiv:1903.06920v127 citations
Originality Incremental advance
AI Analysis

This addresses the challenge of robust super-resolution for clinical imaging, where data quality issues are common, though it appears incremental with specific enhancements to existing GAN frameworks.

The paper tackles the problem of super-resolution in clinical settings where training data often contains corruptions, by proposing novel loss functions that model heavy-tailed non-Gaussian residuals and incorporate manifold-based and perception-based metrics. Results on a large clinical dataset show improvements over state-of-the-art methods.

Super-resolution using deep neural networks typically relies on highly curated training sets that are often unavailable in clinical deployment scenarios. Using loss functions that assume Gaussian-distributed residuals makes the learning sensitive to corruptions in clinical training sets. We propose novel loss functions that are robust to corruptions in training sets by modeling heavy-tailed non-Gaussian distributions on the residuals. We propose a loss based on an autoencoder-based manifold-distance between the super-resolved and high-resolution images, to reproduce realistic textural content in super-resolved images. We propose to learn to super-resolve images to match human perceptions of structure, luminance, and contrast. Results on a large clinical dataset shows the advantages of each of our contributions, where our framework improves over the state of the art.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes