research:skin:epidermis_segmentation_spie2017:epidermis_segmentation_spie2017

This is an old revision of the document!


Automated epidermis segmentation in histopathological images of human skin stained with hematoxylin and eosin

Abstract

Background: Epidermis area is an important observation area for the diagnosis of inflammatory skin diseases and skin cancers. Therefore, in order to develop a computer-aided diagnosis system, segmentation of the epidermis area is usually an essential, initial step. This study presents an automated and robust method for epidermis segmentation in whole slide histopathological images of human skin, stained with hematoxylin and eosin.

Methods: The proposed method performs epidermis segmentation based on the information about shape and distribution of transparent regions in a slide image and information about distribution and concentration of hematoxylin and eosin stains. It utilizes domain-specific knowledge of morphometric and biochemical properties of skin tissue elements to segment the relevant histopathological structures in human skin.

Results: Experimental results on 88 skin histopathological images from three different sources show that the proposed method segments the epidermis with a mean sensitivity of 87%, a mean specificity of 95% and a mean precision of 57%. It is robust to inter- and intra-image variations in both staining and illumination, and makes no assumptions about the type of skin disorder. The proposed method provides a superior performance compared to the existing techniques.

Poster

The poster is available on my Google Drive: PDF (the “Download” button is in the upper right corner of the preview window)

Image datasets (online)

Segmentation results

The following table summarizes segmentation results for each of the compared methods (CET – Haggerty et al., GTSA – Lu et al., THM – Xu et al.).
Statistics of the respective segmentation performance measures are presented here: (click)

:!: Images presented here are low-quality, downscaled versions of the original images!

:!: If there is no segmentation boundary on a given image, than it meas that the respective method failed (not a single pixel has been segmented).

Case ID $I_0$ proposed CET GTSA THM
CMUJ-01_1917687A_1  I0 proposed CET GTSA THM
CMUJ-02_1917687A_2  I0 proposed CET GTSA THM
CMUJ-03_1917687B_1  I0 proposed CET GTSA THM
CMUJ-04_1917687B_2  I0 proposed CET GTSA THM
CMUJ-05_1917733_1  I0 proposed CET GTSA THM
CMUJ-06_1917733_2  I0 proposed CET GTSA THM
CMUJ-07_1917735_1  I0 proposed CET GTSA THM
CMUJ-08_1917735_2  I0 proposed CET GTSA THM
CMUJ-09_1917740_1  I0 proposed CET GTSA THM
CMUJ-10_1917740_2  I0 proposed CET GTSA THM
CMUJ-11_1917742_1  I0 proposed CET GTSA THM
CMUJ-12_1917742_2  I0 proposed CET GTSA THM
CMUJ-13_1917743_1  I0 proposed CET GTSA THM
CMUJ-14_1917743_2  I0 proposed CET GTSA THM
CMUJ-15_1917749_1  I0 proposed CET GTSA THM
CMUJ-16_1917749_2  I0 proposed CET GTSA THM
UBC-01_001a_10  I0 proposed CET GTSA THM
UBC-02_001b_10  I0 proposed CET GTSA THM
UBC-03_001c_10  I0 proposed CET GTSA THM
UBC-04_003a_10  I0 proposed CET GTSA THM
UBC-05_003b_10  I0 proposed CET GTSA THM
UBC-06_003c_10  I0 proposed CET GTSA THM
UBC-07_004_10  I0 proposed CET GTSA THM
UBC-08_005a_10  I0 proposed CET GTSA THM
UBC-09_005b_10  I0 proposed CET GTSA THM
UBC-10_008_10  I0 proposed CET GTSA THM
UBC-11_011a_10  I0 proposed CET GTSA THM
UBC-12_011b_10  I0 proposed CET GTSA THM
UBC-13_011c_10  I0 proposed CET GTSA THM
UBC-14_011d_10  I0 proposed CET GTSA THM
UBC-15_013_10  I0 proposed CET GTSA THM
UBC-16_017a_10  I0 proposed CET GTSA THM
UBC-17_017b_10  I0 proposed CET GTSA THM
UBC-18_018a_10  I0 proposed CET GTSA THM
UBC-19_022_10  I0 proposed CET GTSA THM
UBC-20_023a_10  I0 proposed CET GTSA THM
UBC-21_023b_10  I0 proposed CET GTSA THM
UBC-22_025a_10  I0 proposed CET GTSA THM
UBC-23_025b_10  I0 proposed CET GTSA THM
UBC-24_027_10  I0 proposed CET GTSA THM
UBC-25_029a_10  I0 proposed CET GTSA THM
UBC-26_029b_10  I0 proposed CET GTSA THM
UBC-27_032a_10  I0 proposed CET GTSA THM
UBC-28_032b_10  I0 proposed CET GTSA THM
UBC-29_038_10  I0 proposed CET GTSA THM
UBC-30_043_10  I0 proposed CET GTSA THM
UBC-31_044_10  I0 proposed CET GTSA THM
UBC-32_045a_10  I0 proposed CET GTSA THM
UBC-33_045b_10  I0 proposed CET GTSA THM
UBC-34_052_10  I0 proposed CET GTSA THM
UBC-35_053_10  I0 proposed CET GTSA THM
UBC-36_056a_10  I0 proposed CET GTSA THM
UBC-37_056b_10  I0 proposed CET GTSA THM
UBC-38_057_10  I0 proposed CET GTSA THM
UBC-39_058a_10  I0 proposed CET GTSA THM
UBC-40_058b_10  I0 proposed CET GTSA THM
UBC-41_059_10  I0 proposed CET GTSA THM
UBC-42_061_10  I0 proposed CET GTSA THM
UBC-43_062_10  I0 proposed CET GTSA THM
UBC-44_063_10  I0 proposed CET GTSA THM
UBC-45_064_10  I0 proposed CET GTSA THM
UBC-46_068_10  I0 proposed CET GTSA THM
UBC-47_078a_10  I0 proposed CET GTSA THM
UBC-48_078b_10  I0 proposed CET GTSA THM
UBC-49_078c_10  I0 proposed CET GTSA THM
UBC-50_078d_10  I0 proposed CET GTSA THM
UBC-51_081_10  I0 proposed CET GTSA THM
UBC-52_082_10  I0 proposed CET GTSA THM
UBC-53_085_10  I0 proposed CET GTSA THM
UBC-54_086_10  I0 proposed CET GTSA THM
UBC-55_087a_10  I0 proposed CET GTSA THM
UBC-56_087b_10  I0 proposed CET GTSA THM
UBC-57_088_10  I0 proposed CET GTSA THM
UBC-58_089_10  I0 proposed CET GTSA THM
UBC-59_090_10  I0 proposed CET GTSA THM
UBC-60_094_10  I0 proposed CET GTSA THM
UBC-61_096_10  I0 proposed CET GTSA THM
UBC-62_100a_10  I0 proposed CET GTSA THM
UBC-63_100b_10  I0 proposed CET GTSA THM
UMch-01_32966b  I0 proposed CET GTSA THM
UMch-02_33017b  I0 proposed CET GTSA THM
UMch-03_63220a  I0 proposed CET GTSA THM
UMch-04_63220b  I0 proposed CET GTSA THM
UMch-05_63221a  I0 proposed CET GTSA THM
UMch-06_63221b  I0 proposed CET GTSA THM
UMch-07_Case13s  I0 proposed CET GTSA THM
UMch-08_Case24s  I0 proposed CET GTSA THM
UMch-09_Case26s  I0 proposed CET GTSA THM
UMch-10_Case29s  I0 proposed CET GTSA THM
research/skin/epidermis_segmentation_spie2017/epidermis_segmentation_spie2017.1488632783.txt.gz · Last modified: 2020/03/25 11:46 (external edit)