Novel edge detection method for nuclei segmentation of liver cancer histopathology images

Log-Base2 of Gaussian kernel for nuclei segmentation from colorectal cancer H and E-stained histopathology images

“…Nuclei Segmentation from H and E Stained histopathology images has been one of the most difficult computer vision tasks in recent times [1]. The purpose of nuclei segmentation in any Computer Assisted Diagnosis system is that it simplifies the final classification task of cancer detection [2]. Several researchers utilized conventional image processing techniques like edge detection [2], generalized LoG [3], morphology [4], Active Contour Model [5], etc.…”

“…The purpose of nuclei segmentation in any Computer Assisted Diagnosis system is that it simplifies the final classification task of cancer detection [2]. Several researchers utilized conventional image processing techniques like edge detection [2], generalized LoG [3], morphology [4], Active Contour Model [5], etc. in order to segment only the nuclei portions in histopathology images.…”

“…Therefore, the performance of their methods doesn't work uniformly for any kind of histopathology images. Recently, S. Roy et al [2] have proposed a novel edge detection method, based on computing local standard deviation inside H and E stained histopathology images. The authors claimed that their method avoids noise significantly in their processed image and simultaneously it can extract nuclei information at multi-scale.…”

“…Hence, utilizing handcrafted annotated data (or, readily available ground truth) for nuclei segmentation is not so feasible, to the best of our knowledge. Even under the supervision of pathologists, manually created annotated data may have statistical property far deviated from the original image [2].…”

“…Color normalization and stain separation are other two pre-processing methods which are employed prior to nuclei segmentation method by numerous researchers ( [2], [17]). The main purpose of stain separation method is to separate Hematoxylin and Eosin stain channels, since many researchers claimed that there is a bit overlapping in the color spectrum of H and E stains.…”