Region segmentation method for particle images and apparatus thereof

Abstract

In a region segmentation method for particle images, which photographs stained and non-stained object particles supplied through an image input optical system and which uses at least two of the red-colored image, green-colored image and blue-colored image to discriminate between a background region and an object region where the object particles exist, the region segmentation method includes the processes of: a first process of setting a first group of thresholds in density value for a first group of at least two images selected from the red-colored image, green-colored image and blue-colored image to extract regions greatly differing in density from the background region, and producing a first binary image from the first group of images; a second process of calculating for each pixel a quantity representing the magnitude of change of density in an area neighboring each of the pixels making up each of one or more second group images selected from the red-colored image, green-colored image and blue-colored image, comparing a change of density in the background region with the quantity representing the magnitude of change of density calculated for each pixel, setting a second group of thresholds for extracting regions whose cane of density is larger than tat of the background region, and producing a second binary image from the second group of images; and a third process of performing a logic operation on the first binary image and the second binary image to produce a third binary image showing the background region and the object region.