Spatiotemporally Consistent Indoor Lighting Estimation with Diffusion Priors
This addresses the problem of realistic lighting estimation for computer vision and graphics applications, with incremental improvements in consistency for in-the-wild videos.
The paper tackles the challenge of estimating spatiotemporally varying indoor lighting from a single image or video by proposing a method that uses diffusion priors to optimize a continuous light field represented as an MLP, achieving superior performance over baselines.
Indoor lighting estimation from a single image or video remains a challenge due to its highly ill-posed nature, especially when the lighting condition of the scene varies spatially and temporally. We propose a method that estimates from an input video a continuous light field describing the spatiotemporally varying lighting of the scene. We leverage 2D diffusion priors for optimizing such light field represented as a MLP. To enable zero-shot generalization to in-the-wild scenes, we fine-tune a pre-trained image diffusion model to predict lighting at multiple locations by jointly inpainting multiple chrome balls as light probes. We evaluate our method on indoor lighting estimation from a single image or video and show superior performance over compared baselines. Most importantly, we highlight results on spatiotemporally consistent lighting estimation from in-the-wild videos, which is rarely demonstrated in previous works.