Image Segmentation of Skin Neoplasms Using the Active Contour Method

• Published in: Physics of atomic nuclei 2022-12, Vol.85 (11), p.1956-1960
• Main Authors: Voronin, A. E., Pronichev, A. N., Nikitaev, V. G., Solomatin, M. A., Zanegina, T. P., Arkhangelskaya, I. V., Petukhova, A. I., Bagnova, P. Yu, Soshnina, A. V., Tamrazova, O. B., Sergeev, V. Yu, Sergeev, Yu. Yu
• Format: Article
• Language: English
• Subjects: Analysis; Computer vision; Computers; Decision support systems; Digital imaging; Equipment and supplies; Image processing; Image segmentation; Machine vision; Medical imaging; Medical Physics and Biophysics; Melanoma; Methods; Neoplasms; Particle and Nuclear Physics; Physics; Physics and Astronomy; Skin; Skin tumors; Tumors
• ISSN: 1063-7788; 1562-692X
• DOI: 10.1134/S1063778822090411

Computer vision technologies are being actively introduced into modern life, including medical practice. The development of such technologies has led to the emergence of computer systems that allow the detection and classification of skin diseases with a quality comparable to and in some cases exceeding human capabilities. The article reveals the method of automatic segmentation on the basis of dermatoscopic images provided by doctors obtained using a digital optical device—a dermatoscope. The main goal of the model being developed is to identify the neoplasm zone and areas of hyperpigmentation on images of skin neoplasms for further integration into medical decision support systems for diagnosing of melanoma. As a result of the work carried out, a software package was created that allows segmentation of the neoplasm. As a demonstration of the method, experimental studies of the detection of melanoma boundaries and zones of areas of signs on images of skin neoplasms are presented. The developed system can be used for diagnostic research and educational purposes.