Orientation Sensitive Fuzzy C Means Based Fast Level Set Evolution for Segmentation of Histopathological Images to Detect Skin Cancer
- Publisher:
- Springer International Publishing
- Publication Type:
- Conference Proceeding
- Citation:
- Hybrid Intelligent Systems, 2020, 923, pp. 501-510
- Issue Date:
- 2020-01-01
Closed Access
Filename | Description | Size | |||
---|---|---|---|---|---|
Masood-Al-Jumaily2020_Chapter_OrientationSensitiveFuzzyCMean.pdf | 1.38 MB |
Copyright Clearance Process
- Recently Added
- In Progress
- Closed Access
This item is closed access and not available.
Malignant Melanoma in one of the most deadly skin cancer. In latest cancer diagnosis, pathologists examine biopsies for analysing cell morphology and tissue distribution for making diagnostic assessments. However, this process is quite subjective and has considerable variability. Automated computational diagnostic tools based on quantitative measures can help the diagnosis done by pathologists. The first and foremost step in automated histopathological image analysis is to properly segment the tissue structures such as nest, irregular distribution and melanocytic cells which indicate some disorder or potential cancer. This paper presents a novel technique for automatic segmentation of histopathological skin images, without user intervention. It is based on an innovative approach for utilizing the concepts of clustering and level set evolution. Firstly, the image is pre-processed to enhance the differentiating structural details. Then a novel orientation sensitive Fuzzy C mean clustering is used to generate the initial coarse segmentation and to calculate the controlling parameters for level set evolution. Later, refined fast level set based algorithm is used to finalize the segmentation process. Experimental analysis on a database of 150 histopathological images display the accuracy of the proposed method for detecting the melanocytic areas of the image, with true detection rate of 87.66% and Dice similarity coefficient of 0.88, when segmentation results are compared with images marked by expert pathologists.
Please use this identifier to cite or link to this item: