Weakly supervised semantic segmentation using distinct class specific saliency maps.
Wataru ShimodaKeiji YanaiPublished in: Comput. Vis. Image Underst. (2020)
Keyphrases
- weakly supervised
- saliency map
- superpixels
- input image
- object class
- object segmentation
- natural images
- visual attention
- multiscale
- object recognition
- human visual system
- object detection
- video frames
- viewpoint
- segmentation algorithm
- image patches
- object classes
- appearance model
- bounding box
- higher order
- visual features
- long range
- object categories
- segmentation method
- multiple images
- low level
- graph structure
- partial occlusion
- low level features
- object detectors
- visual information