Multivariate Fields of Experts

AI in healthcare
Published: arXiv: 2508.06490v1
Authors

Stanislas Ducotterd Michael Unser

Abstract

We introduce the multivariate fields of experts, a new framework for the learning of image priors. Our model generalizes existing fields of experts methods by incorporating multivariate potential functions constructed via Moreau envelopes of the $\ell_\infty$-norm. We demonstrate the effectiveness of our proposal across a range of inverse problems that include image denoising, deblurring, compressed-sensing magnetic-resonance imaging, and computed tomography. The proposed approach outperforms comparable univariate models and achieves performance close to that of deep-learning-based regularizers while being significantly faster, requiring fewer parameters, and being trained on substantially fewer data. In addition, our model retains a relatively high level of interpretability due to its structured design.

Paper Summary

Key Innovation
The key innovation of this work is the introduction of multivariate fields of experts (MFoE), a novel framework for learning image priors that generalizes existing fields of experts methods by incorporating multivariate potential functions constructed via Moreau envelopes of the ℓ∞-norm. This allows for improved adaptability across multiple inverse problems.
Practical Impact
The MFoE model has practical impact in several areas, including: * Image denoising: removing noise from images to improve their quality * Image deblurring: removing blur from images to improve their clarity * Compressed-sensing magnetic-resonance imaging (CS-MRI): recovering detailed images from limited data * Computed tomography (CT): reconstructing 3D images from X-ray data The MFoE model outperforms comparable univariate models and achieves performance close to that of deep-learning-based regularizers while being significantly faster, requiring fewer parameters, and being trained on substantially fewer data. Additionally, the model retains a relatively high level of interpretability due to its structured design.
Analogy / Intuitive Explanation
Imagine trying to reconstruct a picture from a bunch of blurry and noisy images. The MFoE model is like a superpowerful image processor that can take these imperfect images and turn them into a clear, high-quality picture. It does this by using a new type of mathematical framework called multivariate fields of experts, which allows it to learn the patterns and structures in the data more effectively than previous methods.
Paper Information
Categories:
eess.IV cs.CV cs.LG eess.SP
Published Date:

arXiv ID:

2508.06490v1

Quick Actions