APPLICATION INTERFACE FOR SEGMENTATION OF LOWER LIMB TELANGIECTASIA

• Grant number: 24/01999-6
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: June 01, 2024
• End date: May 31, 2025
• Field of knowledge: Health Sciences - Medicine - Surgery
• Principal Investigator: Matheus Bertanha
• Grantee: Saulo Augusto Couto
• Host Institution: Faculdade de Medicina (FMB). Universidade Estadual Paulista (UNESP). Campus de Botucatu. Botucatu , SP, Brazil

Abstract

Introduction: Chronic venous disease of the lower limbs has an estimated prevalence of 57.5%, and in its mildest presentation, it is characterized by telangiectasias and reticular veins. To date, the quantification of treatment results is still subjective. The use of artificial intelligence (AI) can be useful in evaluating results more accurately, benefiting both the patient and the doctor. AI will be able to analyze large image banks, using comparisons with a standard process, obtaining results similar to those performed manually by medical professionals. Furthermore, the automation of the method makes it possible for analyses to be carried out at scale. However, the use of AI for this purpose is still not widespread and can be better explored. Objective: To develop an artificial intelligence algorithm for the segmentation and measurement of lower limb telangiectasias using a database of photographic images of people with telangiectasia. Method: Cross-sectional study of methodological accuracy for quantifying telangiectasias. Photographs will be taken in three positions of the leg and thigh of each lower limb for analysis and quantification of telangiectasias. The aim is to sample 240 images with telangiectasias, divided into two groups, with 80% of the images being used for AI training and 20% for method validation. The images will be obtained with camera and luminosity standards, being cropped, resized and treated with the Python programming language to obtain quantification. The binary images will be processed in a photo editor and divided into images and masks for AI training and, later, for methodological validation.