Small Objects Matters in Weakly-supervised Semantic Segmentation.
Cheolhyun MunSanghuk LeeYoungjung UhJunsuk ChoeHyeran ByunPublished in: WACV (2024)
Keyphrases
- weakly supervised
- semantic segmentation
- superpixels
- object classes
- object class
- object detectors
- object segmentation
- topic models
- bounding box
- relation extraction
- pascal voc
- semi supervised
- multiple objects
- object detection
- object categories
- conditional random fields
- scene classification
- segmentation method
- multiscale
- named entities
- image segmentation
- target object
- small number
- d objects
- natural images
- higher order
- multi class
- pairwise
- training set
- object recognition
- training data