FlowIID: Single-Step Intrinsic Image Decomposition via Latent Flow Matching

TL;DR

FlowIID introduces a parameter-efficient, single-step intrinsic image decomposition method using latent flow matching, suitable for real-time applications and competitive with existing models.

Contribution

The paper presents FlowIID, a novel architecture combining VAE-guided latent space with flow matching for efficient, stable, and single-step intrinsic image decomposition.

Findings

FlowIID achieves competitive results across benchmarks.

It is more parameter-efficient than existing models.

FlowIID enables real-time deployment in resource-constrained settings.

Abstract

Intrinsic Image Decomposition (IID) separates an image into albedo and shading components. It is a core step in many real-world applications, such as relighting and material editing. Existing IID models achieve good results, but often use a large number of parameters. This makes them costly to combine with other models in real-world settings. To address this problem, we propose a flow matching-based solution. For this, we design a novel architecture, FlowIID, based on latent flow matching. FlowIID combines a VAE-guided latent space with a flow matching module, enabling a stable decomposition of albedo and shading. FlowIID is not only parameter-efficient, but also produces results in a single inference step. Despite its compact design, FlowIID delivers competitive and superior results compared to existing models across various benchmarks. This makes it well-suited for deployment in resource-constrained and real-time vision applications.