Global and Local Features through Gaussian Mixture Models on Image Semantic Segmentation.
Darwin SaireAdín Ramírez RiveraPublished in: CoRR (2022)
Keyphrases
- semantic segmentation
- gaussian mixture model
- image features
- feature vectors
- low level
- superpixels
- input image
- mixture model
- conditional random fields
- image set
- scene classification
- image retrieval
- test images
- keypoints
- spatial information
- image content
- image representation
- single image
- weakly supervised
- feature extraction
- feature space
- object categories
- multiscale
- image classification
- image regions
- image matching
- feature set
- object classes
- color histogram
- prior knowledge
- segmentation method
- feature descriptors
- bounding box
- bag of words
- object class
- bag of features
- segmentation algorithm
- similarity measure