MORPHOLOGICAL PROCESSING OF IMAGES IN PSEUDOCOLORS AS A METHOD OF CLASSIFICATION IMPROVING

I.K. Vasilyeva, E.A. Pankratova

A method of increasing of image classification reliability that based on morphological filtration of images in pseudocolors is proposed, and applications for whish the use of his is expedient are considered. Algorithm of digital image processing that makes use as input data the results of statistical pixel-by-pixel identification of cells’ classes by the single-channel images of microscopic blood specimens is developed. The finding allows recommending the suggested approach to be used at design of intelligence systems for pattern recognition.

Key words: image processing, classification, topological signatures, production model, knowledgebase.