Video domain adaptation for semantic segmentation using perceptual consistency matching.
Ihsan UllahSion AnMyeongkyun KangPhilip ChikontweHyunki LeeJinwoo ChoiSanghyun ParkPublished in: Neural Networks (2024)
Keyphrases
- domain adaptation
- semantic segmentation
- conditional random fields
- superpixels
- image set
- multiple sources
- labeled data
- semi supervised
- cross domain
- video sequences
- scene classification
- target domain
- unlabeled data
- image matching
- co training
- object classes
- semi supervised learning
- transfer learning
- keypoints
- small number
- prior knowledge
- object categories
- object segmentation
- computer vision
- multiscale
- decision trees