An image-based white blood cell counting method and system

Abstract

The invention relates to an image-based white blood cell counting method and system, wherein the method includes the steps of collecting blood sample images, removing image noise, roughly segmenting cell regions, obtaining the number of cohesive cells and obtaining the total number of white blood cells. Because the present invention adopts the white blood cell automatic segmentation technology of self-adaptive fuzzy clustering technology, the white blood cell segmentation accuracy is higher, and the adhesion cell location segmentation method based on the three-dimensional curved surface can quickly identify the adhesion cells with inconspicuous area and chain code characteristics, adhesion cells The location counting segmentation is more accurate, so that the number of white blood cells can be counted more accurately.

Description

technical field [0001] The invention relates to an image-based white blood cell counting method, and at the same time, the invention also relates to a white blood cell counting system. Background technique [0002] As an indication of certain diseases, the number of white blood cells has important clinical significance for disease diagnosis. Physical method, physical-chemical method and image analysis method are the main methods of white blood cell differential counter at present. Although the automatic counters under these methods can quickly and accurately count white blood cells, they still have the characteristics of bulky equipment, high price, high professionalism in operation, and difficult maintenance, which are not suitable for primary medical and health institutions and mobile medical services. However, there are great limitations in the field of application. For the image-based leukocyte adhesion segmentation and counting method, it mainly focuses on the researc...

AI Technical Summary

Problems solved by technology

In this case, the nucleus area cannot be distinguished on the image, and it is difficult to construct a cell machine model; under a 4-20 magnification microscope, the white blood cells have blurred boundaries and a large number, and the method based on morphological boundary extraction still has low precision. When the adhesion cells cannot be identified, the iterative convergence of the B-snake model is too slow and takes a long time; the boundary of the adhesion cells is blurred, and it is difficult to distinguish the concave and convex points of the boundary, and the discrimination error is large

To sum up, the application of the above methods cannot obtain accurate cell segmentation, cannot completely separate the adherent cells, and there is a problem of low counting accuracy

Images