Saved in:
Bibliographic Details
Main Authors: Zakarin, Daniyar, Wandel, Thiemo, Obukhov, Anton, Dai, Dengxin
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.05000
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915654182895616
author Zakarin, Daniyar
Wandel, Thiemo
Obukhov, Anton
Dai, Dengxin
author_facet Zakarin, Daniyar
Wandel, Thiemo
Obukhov, Anton
Dai, Dengxin
contents We introduce a diffusion-transformer (DiT) framework for single-image reflection removal that leverages the generalization strengths of foundation diffusion models in the restoration setting. Rather than relying on task-specific architectures, we repurpose a pre-trained DiT-based foundation model by conditioning it on reflection-contaminated inputs and guiding it toward clean transmission layers. We systematically analyze existing reflection removal data sources for diversity, scalability, and photorealism. To address the shortage of suitable data, we construct a physically based rendering (PBR) pipeline in Blender, built around the Principled BSDF, to synthesize realistic glass materials and reflection effects. Efficient LoRA-based adaptation of the foundation model, combined with the proposed synthetic data, achieves state-of-the-art performance on in-domain and zero-shot benchmarks. These results demonstrate that pretrained diffusion transformers, when paired with physically grounded data synthesis and efficient adaptation, offer a scalable and high-fidelity solution for reflection removal. Project page: https://hf.co/spaces/huawei-bayerlab/windowseat-reflection-removal-web
format Preprint
id arxiv_https___arxiv_org_abs_2512_05000
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Reflection Removal through Efficient Adaptation of Diffusion Transformers
Zakarin, Daniyar
Wandel, Thiemo
Obukhov, Anton
Dai, Dengxin
Computer Vision and Pattern Recognition
Artificial Intelligence
We introduce a diffusion-transformer (DiT) framework for single-image reflection removal that leverages the generalization strengths of foundation diffusion models in the restoration setting. Rather than relying on task-specific architectures, we repurpose a pre-trained DiT-based foundation model by conditioning it on reflection-contaminated inputs and guiding it toward clean transmission layers. We systematically analyze existing reflection removal data sources for diversity, scalability, and photorealism. To address the shortage of suitable data, we construct a physically based rendering (PBR) pipeline in Blender, built around the Principled BSDF, to synthesize realistic glass materials and reflection effects. Efficient LoRA-based adaptation of the foundation model, combined with the proposed synthetic data, achieves state-of-the-art performance on in-domain and zero-shot benchmarks. These results demonstrate that pretrained diffusion transformers, when paired with physically grounded data synthesis and efficient adaptation, offer a scalable and high-fidelity solution for reflection removal. Project page: https://hf.co/spaces/huawei-bayerlab/windowseat-reflection-removal-web
title Reflection Removal through Efficient Adaptation of Diffusion Transformers
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2512.05000