Model for Estimating the Morphological Characteristics of Structureless Areas of Pigmented Skin Neoplasms

Abstract The problem of early diagnosis of skin melanoma by dermatoscopic images of skin neoplasms is considered. A model for assessing the diagnostically significant morphological characteristics of neoplasm elements—structureless areas—is proposed. An experimental study has been conducted to determine the adequacy of the model. The most informative features for assessment of morphological characteristics have been identified. The best classification accuracy (94%) has been achieved for feature A1. The results of this study can be applied in the development of medical decision support systems for the diagnosis of skin melanoma.

Authors
Tamrazova O.B. 1 , Nikitaev V.G. , Pronichev A.N. , Sergeev V.Yu. , Sergeev Yu.Yu. , Lim A.O. , Kozlov V.S.
Number of issue
11
Pages
1943-1947
Status
Published
Volume
85
Year
2022
Organizations
  • 1 Peoples Friendship University of Russia
Keywords
recognition; segmentation; digital image processing; dermatoscopy; diagnosis of melanoma; oncological diagnostics; computer vision
Date of creation
21.04.2023
Date of change
21.04.2023
Short link
https://repository.rudn.ru/en/records/article/record/93446/
Share

Other records