CGMGM: A Cross-Gaussian Mixture Generative Model for Few-Shot Semantic Segmentation.
Junao ShenKun KuangJiaheng WangXinyu WangTian FengWei ZhangPublished in: AAAI (2024)
Keyphrases
- generative model
- gaussian mixture
- em algorithm
- mixture model
- expectation maximization
- object categories
- conditional random fields
- superpixels
- gaussian mixture model
- probabilistic model
- weakly supervised
- maximum likelihood
- topic models
- bayesian framework
- probability density function
- visual features
- object classes
- object class
- closed form
- semi supervised
- video sequences
- scene classification
- graphical models
- prior knowledge
- maximum a posteriori
- unsupervised learning
- information retrieval
- image classification
- ground truth
- low level
- object recognition
- multiscale
- computer vision