Unsupervised Anomaly Instance Segmentation for Baggage Threat Recognition.
Taimur HassanSamet AkcayMohammed BennamounSalman H. KhanNaoufel WerghiPublished in: CoRR (2021)
Keyphrases
- unsupervised learning
- image parsing
- fully unsupervised
- recognition rate
- correct recognition
- numeral strings
- bayesian image segmentation
- text segmentation
- object recognition
- level set
- medical images
- segmentation of textured images
- image segmentation
- recognition process
- segmentation algorithm
- recognition accuracy
- symbol recognition
- feature extraction and matching
- automatic recognition
- segmentation of natural images
- activity recognition
- segmentation method
- feature extraction
- multiscale
- pattern recognition
- character segmentation
- features extraction
- texture segmentation
- region growing
- shape prior
- intrusion detection
- segmentation accuracy
- semi supervised
- computer vision
- action recognition
- image processing
- fully automatic
- image analysis
- object detection
- supervised training
- object segmentation
- supervised learning
- gesture recognition
- anomaly detection
- segmented images