Serviços Personalizados
Journal
Artigo
Indicadores
- Citado por SciELO
- Acessos
Links relacionados
- Similares em SciELO
Compartilhar
Nova scientia
versão On-line ISSN 2007-0705
Resumo
PUENTE RAMIREZ, N. P.; TORRES-TREVINO, L. M. e SANCHEZ-CANTU, J. E.. Computational modeling by linear regression and symbolic α β to optically determine the pore size of the skin. Nova scientia [online]. 2015, vol.7, n.14, pp.218-235. ISSN 2007-0705.
The appearance of pores of the skin is normally caused by overactive sebaceous glands, however, in recent years it has been reported that in addition is affected by various diseases including diabetes 1, obesity 2 or cancer 3. Generally the main factors contributing to the enlarged pores are age, male gender, acne, chronic sun exposure and genetic predisposition. The development of a noninvasive method to know the conditions of the skin could help to relate the factors that determine the pore size of the skin. This paper aims to characterize, through a computational model and with the help of an optical scheme, the pore size of the skin. The physical principle is to illuminate with LEDs (543nm) the forearm of volunteers, the scattering of light generated by the skin is collected by a CCD camera. The image acquisition, processing and statistical analysis are part of the methodology for data adquisition. The mathematical model proposed relates gender, age, skin tone and pore size (calculated statistically on images acquired optically). A database is generated and used to build mathematical models by symbolic regression with Particle Swarm Optimization (PSO), and a comparison by linear regression is performed. Some statistical indicators such as mean square, error prediction, error sum of squares and percentage of variability are used in the comparison. The results indicate that the proposed model, for the pore size of each individual, can help make an objective interpretation of disease indicators that directly affect the pore size in the skin.
Palavras-chave : skin pore; linear regression; PSO; numerical modeling.