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

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Singla, Mithlesh, Kumari, Seema, Raman, Shanmuganathan
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917209288212480
author Singla, Mithlesh
Kumari, Seema
Raman, Shanmuganathan
author_facet Singla, Mithlesh
Kumari, Seema
Raman, Shanmuganathan
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2601_12329
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle FlowIID: Single-Step Intrinsic Image Decomposition via Latent Flow Matching
Singla, Mithlesh
Kumari, Seema
Raman, Shanmuganathan
Computer Vision and Pattern Recognition
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.
title FlowIID: Single-Step Intrinsic Image Decomposition via Latent Flow Matching
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2601.12329