Subspace Vector Quantization and Markov Modeling for Cell Phase Classification

Vector quantization (VQ) and Markov modeling methods for cellular phase classification using time-lapse fluorescence microscopic image sequences have been proposed in our previous work. However the VQ method is not always effective because cell features are treated equally although their importance may not be the same. We propose a subspace VQ method to overcome this drawback. The proposed method can automatically weight cell features based on their importance in modeling. Two weighting algorithms based on fuzzy c-means and fuzzy entropy clustering are proposed. Experimental results show that the proposed method can improve the cell phase classification rates.