Theapplied Research Of Image Fusion And Contour Detection In The Screening Of Liquid-base D Cervical Cells

The following problems in the process of cervical cancer pathology investigation(1) microscope field of view is relatively small, only display a part of the information ofcervical cell biopsy in single vision;(2) part of the cervical nucleus edge is weak inthe microscopic images;(3) cervical cells can not be completely clear and focused andfully expressed information of cells in the one image. In this paper, do some researchin the problems in the process of cervical cancer pathology investigation and usedifferent of image processing algorithms to solve these problems.(1) Full expressing the complete information of cells in the one image is the firstknow the conditions of cell image recognition in the screening of liquid-basedcervical cells (especially suspected variant or mutant cells). A fusion algorithm ofmulti-focus image sequence of cervical cell is represented in this paper. Use imagefusion technology to preserve the original information and integrate the clearest areainto the one image. Experimental results can be seen that the fusion image is smooth,visual effect, effectively preserve the original information, and no obvious blockeffect and distortion(2) Microscope field of view is relatively small, only display a part of theinformation in the cervical cell biopsy in single vision, can not display the completeand full cell slice information. The harris corner and inter-related criteria-based imageregistration method is introduced in liquid-based cervical cell image registration. First,use the harris corner and inter-related criteria to finish image registration roughly.Second, use the information of image edge to finish image registration accurately.Experimental results can be seen that a better result can be obtained by this method.(3) A problem that can not accurately detect the weak edge of liquid-basedcervical nucleus in the screening of cells. Experimental results can be seen that theliquid-based cervical nucleus edge is basically the same as what the human eye see.In this paper, do some applied research in image fusion and contour detection tosolve some problems in the screening of liquid-based cervical cells and achieve thedesired purpose.