Material Editing Using a Physically Based Rendering Network
This addresses a challenging task for content creators by enabling material editing in images through a physically based approach.
The paper tackles the problem of editing object materials in images by proposing an end-to-end network that predicts intrinsic properties like shape, illumination, and material, then uses a differentiable rendering layer to synthesize images, enabling physically plausible material edits.
The ability to edit materials of objects in images is desirable by many content creators. However, this is an extremely challenging task as it requires to disentangle intrinsic physical properties of an image. We propose an end-to-end network architecture that replicates the forward image formation process to accomplish this task. Specifically, given a single image, the network first predicts intrinsic properties, i.e. shape, illumination, and material, which are then provided to a rendering layer. This layer performs in-network image synthesis, thereby enabling the network to understand the physics behind the image formation process. The proposed rendering layer is fully differentiable, supports both diffuse and specular materials, and thus can be applicable in a variety of problem settings. We demonstrate a rich set of visually plausible material editing examples and provide an extensive comparative study.