Fig. 1: Details of our proposed dataset (HEADSET) in terms of modalities and use cases. (A): Textured 3D meshes with 6 facial expressions, (B): three types of 3D point cloud with participant wearing an HMD, (C): RGB images and depth maps from 3 views, (D): multi-view representation of LF, (E): point clouds evaluation results, (F): result of HMD removal, and (G): classified LF images with 6 facial expressions.