Segmentation on Micrograph of Blood Cells Based on Improved GAC Model
Bin B. Zhu · Jisuanji fangzhen · 2011
To shoot drawbacks of current typical algorithms used for segmentation on micrograph of blood cells,this paper proposes an improved Geodesic Active Contour Model(GAC) on aspects of filtering and the contrast parameter.In the process of segmentation,an initial contour,acquired by combination of non-differential SUSAN and morphologic operations,is provided to the improved GAC model which evolves to reach refined true cell edges,at last,with distance transform to search centers of overlapped cells,boundaries between overlapped cells are found by geodesic dilation.The simulation indicates that the proposed algorithm acquires more integral area of cells and continuous smooth edges,and favorable insistence to noise disturbance.