Extraction method of glandular areas in prostate biopsy image

Number of prostate cancer patients is increasing. So system of supporting diagnosis is needed by pathologists. Biopsy images stained by HE staining, but the state depends on pathologist, individual condition, and environments. In this study, we try to robust image segmentation method to divide glandular into components. Our proposed method is to use fixed color range for background, Otsu method with Laplacian histosgram for nuclei, and not only color information but also texture features for interstitium and cytoplasm.