We proposed a neural network approach for microfacet material energy compensation. Our method only takes roughness and F0 parameters for GGX model and predicts energy compensated BRDF values. The model is very effective in the inference stage, and can handle isotropic/anisotropic, colored materials.

We proposed a decouple pipeline, using a neural network as a coordinate-independent feature extractor and a light-weight decision tree as a coordinate-dependent regressor, to relocate camera 6D pose by a RGB image. Our approach can be adapted in a new scene with only few-shot data while achieving reasonable relocalization accuracy.