Reflection Generation for Composite Image Using Diffusion Model

2026-04-02

AI Summary

Existing image synthesis methods predominantly focus on shadows while largely neglecting reflections, and lack effective mechanisms for generating environment-consistent reflections. This work addresses this gap by introducing DEROBA, the first large-scale object reflection dataset, and proposing a type-aware diffusion model framework that explicitly distinguishes between specular and diffuse reflections through the injection of reflection location and appearance priors. By incorporating these physically informed cues, the method achieves significantly improved performance in both physical consistency and visual realism compared to existing approaches. The proposed framework establishes a new benchmark for reflection generation and fills a critical void in the literature on photorealistic scene synthesis.

Abstract

Image composition involves inserting a foreground object into the background while synthesizing environment-consistent effects such as shadows and reflections. Although shadow generation has been extensively studied, reflection generation remains largely underexplored. In this work, we focus on reflection generation. We inject the prior information of reflection placement and reflection appearance into foundation diffusion model. We also divide reflections into two types and adopt type-aware model design. To support training, we construct the first large-scale object reflection dataset DEROBA. Experiments demonstrate that our method generates reflections that are physically coherent and visually realistic, establishing a new benchmark for reflection generation.

Problem

• reflection generation
• image composition
• environment-consistent effects
• diffusion model

Innovation

• reflection generation
• diffusion model
• type-aware modeling
• DEROBA dataset
• image composition