AUTOMATIC CLASSIFYING METHOD FOR RED BLOOD CORPUSCLE MICRONUCLEUS CELL AND AUTOMATIC CLASSIFYING DEVICE FOR RED BLOOD CORPUSCLE MICRONUCLEUS CELL AND RED BLOOD CORPUSCLE MICRONUCLEUS CELL SAMPLE GENERATING METHOD

摘要

PURPOSE:To automatize a micronucleus test by generating a density histogram of each image data of a red blood corpuscle area, further dividing a micronucleus part, and a reticulate granular substance part into areas and calculating a feature parameter, calculating a feature parameter related to a cell shape from the image data, as well, executing a branched logical operation, based thereon and executing the classification. CONSTITUTION:A feature extraction processor 8 connected to an image memory 7 is provided with a function for generating image data, a function for dividing a red blood corpuscle and a background into areas, a function for generating a density histogram, and a function for calculating a feature parameter, and a recognition processor 9 applies branched logic to the feature parameter calculated by the feature extraction processor 8 and classifies the red blood corpuscle. Accordingly, the feature parameter and a decision reference value can be allowed to correspond to one-to-one. In such a way, the red blood corpuscle having a micronucleus can be classified automatically with the same accuracy as the visual observation classification.