Unsupervised Crack Segmentation with Candidate Crack Region Identification and Graph Neural Network Clustering.
Hein Thura AungWuttipong KumwilaisakPublished in: IAIT (2023)
Keyphrases
- background subtraction
- neural network
- foreground and background
- unsupervised learning
- normalized cut
- self organizing maps
- graph theoretic
- agglomerative clustering
- clustering algorithm
- graph partitioning
- graph clustering
- road surface
- information bottleneck
- k means
- clustering method
- unsupervised classification
- unsupervised feature selection
- unsupervised manner
- optimal segmentation
- som neural network
- region of interest
- graph model
- image regions
- random walk
- homogeneous regions
- information theoretic
- segmented images
- unsupervised clustering
- region segmentation
- graph structure
- image segmentation
- spectral clustering
- data clustering
- connected components
- segmentation method
- semi supervised
- supervised learning
- medical images
- min cut max flow
- multiscale
- cluster validation
- artificial neural networks
- edge detection
- text segmentation
- segmented regions
- weighted graph
- fully unsupervised
- level set
- back propagation
- energy function
- grey level
- region growing