Intrinsic Image

Intrinsic image decomposition aims to separate a photograph into its underlying components, primarily albedo (surface reflectance) and shading (illumination), providing a more physically meaningful representation of a scene. Current research focuses on improving the accuracy and robustness of this decomposition using deep learning models, particularly diffusion models and generative adversarial networks (GANs), often incorporating physically-based priors or illumination-invariant features to address the ill-posed nature of the problem. This research is significant for applications in image editing, compositing, rendering, and computer vision tasks requiring a deeper understanding of scene geometry and material properties.