Automated segmentation of abnormal cervical cells using global and local graph cuts

In this paper, a global and local scheme based on graph cuts approach is proposed to segment cervical cells in images with a mix of healthy and abnormal cells. For cytoplasm segmentation, on the A* channel enhanced image, the multi-way graph cut is performed globally, which can effectively extract cytoplasm boundaries when image histograms present non-bimodal distribution. For nucleus especially abnormal nucleus segmentation, we propose to use graph cut adaptively and locally, which allows the combination of intensity, texture, boundary and region information together. Two concave-based approaches are integrated to split the touching-nuclei. On 21 cervical cell images with non-ideal imaging condition and pathology, our segmentation method achieved a 93% accuracy for cytoplasm, and a 88.4% F-measure for abnormal nuclei, both outperformed state of the art works in terms of accuracy.