Robust clustering based on a maximum-likelihood method for estimating a suitablenumber of clusters.
Naoyuki IchimuraPublished in: Systems and Computers in Japan (1997)
Keyphrases
- maximum likelihood
- computationally efficient
- high accuracy
- synthetic data
- experimental evaluation
- detection method
- gaussian distribution
- cost function
- computational cost
- input data
- em algorithm
- unsupervised clustering
- image segmentation
- highly accurate
- fuzzy clustering
- image matching
- high precision
- cross validation
- prior knowledge
- pairwise
- similarity measure