Motion quaternion-based motion estimation method of MYO using K-means algorithm and Bayesian probability.
Yunsick SungHaitao GuoSang-Geol LeePublished in: Soft Comput. (2018)
Keyphrases
- k means
- motion estimation
- clustering method
- cost function
- detection algorithm
- block matching
- dynamic programming
- high accuracy
- preprocessing
- optimization algorithm
- computational complexity
- computational cost
- recognition algorithm
- objective function
- matching algorithm
- optical flow
- input data
- hierarchical clustering
- segmentation algorithm
- phase correlation
- initial cluster centers
- prediction error
- motion estimation algorithm
- probabilistic model
- estimation algorithm
- segmentation method
- learning algorithm
- block matching algorithm
- global motion
- similarity measure
- cluster centers
- image sequences
- optimal solution
- clustering algorithm
- bayesian formulation
- model selection
- matching process
- reference frame
- data clustering
- expectation maximization