Dermoscopic image segmentation combining pigment separation and deep learning networks

Yongzhong Zhang; Hexiao Huang; Junzhu Zhang; Yan Ma

doi:10.1117/12.2671702

21 March 2023 Dermoscopic image segmentation combining pigment separation and deep learning networks

Yongzhong Zhang, Hexiao Huang, Junzhu Zhang, Yan Ma

Proceedings Volume 12609, International Conference on Computer Application and Information Security (ICCAIS 2022); 1260910 (2023) https://doi.org/10.1117/12.2671702
Event: International Conference on Computer Application and Information Security (ICCAIS 2022), 2022, ONLINE, ONLINE

Abstract

Dermoscopic image segmentation aims to detect damaged skin areas. It is difficult to detect skin lesions accurately due to factors such as the method of image acquisition, the characteristics of the lesion, and the texture of the skin. This paper proposes a dermatoscopic image segmentation algorithm that combines the pigment separation algorithm with the Segnet network model. First, pigment separation is performed on the dermatoscopic images to obtain the corresponding melanin and hemosiderin images. Afterwards, the melanin and hemosiderin images are converted to single-channel grayscale maps and merged with the original images to produce image data with a channel number of 5. The channel-expanded images are then segmented using Segnet deep neural network. Experimental results on the ISIC-2018 dermoscopy image dataset show that the proposed algorithm achieves better segmentation results in terms of accuracy, sensitivity, and specificity.

Citation Download Citation

Yongzhong Zhang, Hexiao Huang, Junzhu Zhang, and Yan Ma "Dermoscopic image segmentation combining pigment separation and deep learning networks", Proc. SPIE 12609, International Conference on Computer Application and Information Security (ICCAIS 2022), 1260910 (21 March 2023); https://doi.org/10.1117/12.2671702

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available