An SVM based scoring evaluation system for fluorescence microscopic image classification

This paper proposes a scoring evaluation system to the fluorescence microscopic image classification based on the support vector machine (SVM). We define the similarity scores for each testing sample based on its relative distance to the SVM separating hyperplanes and the training clustering centers in the feature space. The method proposed calculates similarity scores through a two-stage process that converts the SVM´s classification results to a quantitative description. The scores can precisely reflect how similar a testing sample to all the categories and provide a reference to further investigation of fluorescence microscopic images.