Bringing Background into the Foreground: Making All Classes Equal in Weakly-Supervised Video Semantic Segmentation.
Fatemehsadat SalehMohammad Sadegh Ali AkbarianMathieu SalzmannLars PeterssonJose M. AlvarezPublished in: ICCV (2017)
Keyphrases
- weakly supervised
- semantic segmentation
- foreground objects
- superpixels
- video sequences
- bounding box
- object class
- object classes
- object segmentation
- video frames
- relation extraction
- object detectors
- semi supervised
- multiscale
- natural images
- object categories
- object recognition
- topic models
- conditional random fields
- input image
- named entities
- segmentation method
- segmentation algorithm
- long range
- scene classification
- object detection
- viewpoint
- image segmentation
- active learning
- training set
- image sequences