ViSER: Video Surface Embeddings for Articulated 3D Shape Reconstruction

Results on athletic human

DAVIS-parkour. Top left: reference video. Top middle: reconstructed nonrigid 3D shape. Top right: reconstruction 3D shape at 1st frame. Bottom left: trajectory. Bottom middle: part segmnetation. Bottom right: textured mesh.

DAVIS-dance. Top left: reference video. Top middle: reconstructed nonrigid 3D shape. Top right: reconstruction 3D shape at 1st frame. Bottom left: trajectory. Bottom middle: part segmnetation. Bottom right: textured mesh.

DAVIS-breakdance-flare. Top left: reference video. Top middle: reconstructed nonrigid 3D shape. Top right: reconstruction 3D shape at 1st frame. Bottom left: trajectory. Bottom middle: part segmnetation. Bottom right: textured mesh.

DAVIS-breakdance. Top left: reference video. Top middle: reconstructed nonrigid 3D shape. Top right: reconstruction 3D shape at 1st frame. Bottom left: trajectory. Bottom middle: part segmnetation. Bottom right: textured mesh.

Ballet-1. Top left: reference video. Top middle: reconstructed nonrigid 3D shape. Top right: reconstruction 3D shape at 1st frame. Bottom left: trajectory. Bottom middle: part segmnetation. Bottom right: textured mesh.

Ballet-2. Top left: reference video. Top middle: reconstructed nonrigid 3D shape. Top right: reconstruction 3D shape at 1st frame. Bottom left: trajectory. Bottom middle: part segmnetation. Bottom right: textured mesh.

Ballet-3. Top left: reference video. Top middle: reconstructed nonrigid 3D shape. Top right: reconstruction 3D shape at 1st frame. Bottom left: trajectory. Bottom middle: part segmnetation. Bottom right: textured mesh.