uroboros: Cycle Consistent Diffusion-based Models for Forward and Inverse Rendering
ICCV 2025
TL;DR
Ouroboros consists of two single-step diffusion models that handle both forward and inverse rendering with mutual reinforcement. our approach extends intrinsic decomposition to both indoor and outdoor scenes and introduces a cycle consistency mechanism that ensures coherence between forward and inverse rendering outputs.
Method Overview
Overview of our framework pipeline. (a) presents the training pipeline of our single-step Diffusion-based inverse and forward rendering model. For inverse rendering, the model takes the image I and text prompt indicating the output intrinsic maps as input to finetune the latent diffusion UNet. For forward rendering, the model is fed with concatenated intrinsic maps along with simple image description to estimate the original image. (b) provides the overview of cycle training pipeline.
Video Inference Overview
Overlapping windows are processed sequentially, with latent representations from previous windows guiding the initialization of overlapping regions. In practice, the window size and overlap are larger than the figure shown.
Video Comparison