Disentangling spatio-temporal knowledge for weakly supervised object detection and segmentation in surgical video.
Guiqiu LiaoMatjaz JoganSai KoushikEric EatonDaniel A. HashimotoPublished in: CoRR (2024)
Keyphrases
- weakly supervised
- superpixels
- object class
- topic models
- semi supervised
- temporal information
- relation extraction
- video data
- video sequences
- object detectors
- named entities
- object segmentation
- multiscale
- bayesian networks
- object detection
- co occurrence
- image processing
- video frames
- graph cuts
- input image
- prior knowledge
- object recognition
- image sequences