Unifying Deep Stochastic Processes for Image Enhancement

TL;DR

This paper unifies various deep stochastic processes for image enhancement under a common SDE framework, clarifying their differences and enabling fair comparison and rapid prototyping.

Contribution

It introduces a unified perspective on stochastic image enhancement methods, classifies them into three families, and releases a modular library for experimentation.

Findings

No single method consistently outperforms others across tasks.

Design choices like drift, diffusion, and boundary conditions significantly impact performance.

The unified framework facilitates fair comparison and understanding of stochastic image enhancement methods.

Abstract

Deep stochastic processes have recently become a central paradigm for image enhancement, with many methods explicitly conditioning the stochastic trajectory on the degraded input. However, the relationship between these conditional processes and standard diffusion models remains unclear. In this work, we introduce a unified perspective on stochastic image enhancement by classifying recent methods into three families of continuous-time processes: unconditional diffusion models, Ornstein-Uhlenbeck (OU) processes, and diffusion bridges. We show that all of these approaches arise from a common stochastic differential equation (SDE) formulation. This framework makes explicit that seemingly disparate methods differ primarily in their drift and diffusion terms, terminal distributions, and boundary conditions, while schedulers and samplers constitute orthogonal design choices. Leveraging this unification, we conduct a controlled empirical study across multiple image enhancement tasks using identical architectures and training protocols. Our results reveal no consistently dominant method; instead, we identify and disentangle the specific design choices that most strongly influence performance. Finally, we release ItoVision, a modular PyTorch library that implements the unified framework and enables rapid prototyping and fair comparison of stochastic image enhancement methods.