Automated skin lesion segmentation of dermoscopic images using GrabCut and k-means algorithms

• Published in: IET computer vision Vol. 12; no. 8; pp. 1088 - 1095
• Main Authors: Jaisakthi, Seetharani Murugaiyan, Mirunalini, Palaniappan, Aravindan, Chandrabose
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 01.12.2018; Wiley
• Subjects: automated skin lesion diagnosis system; automatic skin lesion segmentation method; biomedical optical imaging; cancer; colour features; dermoscopic images; dermoscopic imaging system; diagnostic capability enhancement; filtering techniques; GrabCut segmentation algorithm; graph theory; image classification; image colour analysis; image filtering; image segmentation; ISIC 2017 challenge dataset; k-means clustering algorithm; lesion region; medical image processing; object detection; pattern clustering; PH2 dataset; preprocessing step; segmentation step; skin; skin cancer; skin lesion classification; skin lesion detection; Special Issue: Computer Vision in Cancer Data Analysis; visual examination; dermoscopic imaging system; skin cancer; image classification; graph theory; skin; ISIC 2017 challenge dataset; image filtering; GrabCut segmentation algorithm; k-means clustering algorithm; object detection; lesion region; filtering techniques; image segmentation; skin lesion detection; colour features; image colour analysis; medical image processing; segmentation step; diagnostic capability enhancement; automated skin lesion diagnosis system; visual examination; preprocessing step; pattern clustering; dermoscopic images; cancer; PH2 dataset; automatic skin lesion segmentation method; skin lesion classification; biomedical optical imaging
• ISSN: 1751-9632; 1751-9640; 1751-9640
• DOI: 10.1049/iet-cvi.2018.5289

Skin cancer is the most common type of cancer in the world and the incidents of skin cancer have been rising over the past decade. Even with a dermoscopic imaging system, which magnifies the lesion region, detecting and classifying skin lesions by visual examination is laborious due to the complex structures of the lesions. This necessitates the need for an automated skin lesion diagnosis system to enhance the diagnostic capability of dermatologists. In this study, the authors propose an automatic skin lesion segmentation method which can be used as a preliminary step for lesion classification. The proposed method comprises two major steps, namely preprocessing and segmentation. In the preprocessing step, noise such as illumination, hair and rulers are removed using filtering techniques and in the segmentation phase, skin lesions are segmented using the GrabCut segmentation algorithm. The k-means clustering algorithm is then used along with the colour features learnt from the training images to improve the boundaries of the segments. To evaluate the authors’ proposed method, they have used ISIC 2017 challenge dataset and ${\rm P}{\rm H}^2$PH2 dataset. They have obtained Dice coefficient values of 0.8236 and 0.9139 for ISIC 2017 test dataset and ${\rm P}{\rm H}^2$PH2 dataset, respectively.