A Weakly-Supervised Semantic Segmentation Approach Based on the Centroid Loss: Application to Quality Control and Inspection.
Kai YaoAlberto OrtizFrancisco Bonnín-PascualPublished in: IEEE Access (2021)
Keyphrases
- quality control
- weakly supervised
- semantic segmentation
- superpixels
- topic models
- object class
- scene classification
- conditional random fields
- semi supervised
- relation extraction
- object classes
- object categories
- named entities
- computer vision
- image processing
- object segmentation
- segmentation algorithm
- natural images
- domain specific
- text mining
- training set
- pascal voc
- multiscale