Unsupervised Disentanglement of Pose, Appearance and Background from Images and Videos.
Aysegul DundarKevin J. ShihAnimesh GargRobert PottorfAndrew TaoBryan CatanzaroPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2022)
Keyphrases
- input image
- bounding box
- image features
- viewing angle
- image retrieval
- static images
- visual appearance
- visual patterns
- pose and lighting
- three dimensional
- calibrated cameras
- geometric information
- position and orientation
- probe image
- image classification
- improve recognition accuracy
- object appearance
- object representation
- complex background
- pixel wise
- image set
- object recognition
- camera pose
- illumination variations
- partial occlusion
- camera viewpoint
- segmentation algorithm
- d objects