Microscopic Image Of Biological Microscopic Cell Image Segmentation Based On Retinex Theory And Clustering Approach

Image segmentation is an important problem in image processing. Up to now, there is still no common ways or standardized methods for image segmentation. Cell microscopic image segmentation is an important pre-basis in medical statistics for number of cells and image parameters of each cell including area, perimeter and volume. The quality of image segmentation will have direct influence on the subsequent processing and analysis of cell parameters.This article summarized the current methods for image segmentation. I suggested a cell segmentation approach based on the principle of Retinex theory on VC++6.0+opencv platform and realized the image segmentation of cell microscopic image. This approach segments image directly after isolating a certain percentage of the high-frequency images, or combining with the region growing image segmentation method. This approach firstly segments image based on the principle of Retinex theory. After corrosion, expansion, median filtering to remove noise while retention characteristics of cell edge, and filling the interior contour, the original image was filled to the final image in the segmentation template. This method has good segmentation effect.Meanwhile, this article also studied the background segmentation of cell microscopic image using K-means clustering approach. Combined with adaptive edge detection, corrosion expansion, median filtering, filling the interior contour and other image processing techniques to extract the cell region and realized the algorithm.K-means clustering segmentation in efficiency is better than the region growing method. However, the regional growth method is more flexible than K-means clustering segmentation in selectivity. These two segmentation methods both have good segmentation effect, providing a good foundation for the subsequent image processing.