Extraction of Leukocyte Areas Based on Improved ITTI and Particle Swarm Optimization Algorithm

In the process of pathological analysis of the Leukocyte microscopic image, the leukocyte areas are regions of interest(ROI). The ITTI visual model is an effective method of extracting the ROI from image. For improving the extracting accuracy, an object extracting method combining the improved ITTI visual model with particle swarm optimization algorithm is proposed and used to extract the ROI from the bone marrow cell image. Firstly, based on Gaussian filter and multi-scale normalization, features of the orientation, brightness, and color are computed from the original image. And then, according the fact that the sensitivity of eyes is not the same as different features the saliency map is obtained with adaptive coefficient from three significant characteristics. Finally, by using Otsu method based on the improved particle swarm optimization(PSO) algorithm, the ROIs are extracted and subsequent processed with method of morphology. Experimental results show that this method can extract white blood cell areas perfectly, which is helpful for improving the efficiency of pathological analysis.