Segmentation of pathology microscopic images

The light microscopic analysis of the number and shape of cells in pathology is important for the diagnosis and assessment of clinical behavior of disease conditions. The fundamental step of this work is to separate the cells from the background. To segment objects from such uneven background images, a fixed threshold is not suitable. The authors have proposed a new adaptive thresholding method using variational theory. In this paper, this method is introduced to segment pathological images under the light microscope. The comparison of the segmentation results of the authors' method and Otsu's (1979) thresholding method shows the advantage of the authors' method.