This paper describes a software design for a preprocessing and a feature extraction of an automatic white blood cell differential counting system.
The digital image data are transferred to HP3000 mini computer through Z-2 micro computer. The digital image data are proceeding the preprocessing, the feature extraction, the classifier and the counting.
In preprocessing, the digital image data are divided the three areas of nucleus, cytoplasm and background by using the difined octal codes and octal code transforms.
Classification accuracy of 95 percent was obtained by minimum distance classifier using the best 11 features.