J Y Sun1, S W Kim1, S H Lee1, J E Choi2, S J Ko1. 1. Department of Electrical Engineering, Korea University, Seoul, Korea. 2. Department of Dermatology, College of Medicine, Korea University, Seoul, Korea.
Abstract
BACKGROUND/ PURPOSE: As facial pore widening and its treatments have become common concerns in the beauty care field, the necessity for an objective pore-analyzing system has been increased. Conventional apparatuses lack in usability requiring strong light sources and a cumbersome photographing process, and they often yield unsatisfactory analysis results. This study was conducted to develop an image processing technique for automatic facial pore analysis. METHODS: The proposed method detects facial pores using multi-scale detection and optimal scale selection scheme and then extracts pore-related features such as total area, average size, depth, and the number of pores. Facial photographs of 50 subjects were graded by two expert dermatologists, and correlation analyses between the features and clinical grading were conducted. We also compared our analysis result with those of conventional pore-analyzing devices. RESULTS: The number of large pores and the average pore size were highly correlated with the severity of pore enlargement. In comparison with the conventional devices, the proposed analysis system achieved better performance showing stronger correlation with the clinical grading. CONCLUSION: The proposed system is highly accurate and reliable for measuring the severity of skin pore enlargement. It can be suitably used for objective assessment of the pore tightening treatments.
BACKGROUND/ PURPOSE: As facial pore widening and its treatments have become common concerns in the beauty care field, the necessity for an objective pore-analyzing system has been increased. Conventional apparatuses lack in usability requiring strong light sources and a cumbersome photographing process, and they often yield unsatisfactory analysis results. This study was conducted to develop an image processing technique for automatic facial pore analysis. METHODS: The proposed method detects facial pores using multi-scale detection and optimal scale selection scheme and then extracts pore-related features such as total area, average size, depth, and the number of pores. Facial photographs of 50 subjects were graded by two expert dermatologists, and correlation analyses between the features and clinical grading were conducted. We also compared our analysis result with those of conventional pore-analyzing devices. RESULTS: The number of large pores and the average pore size were highly correlated with the severity of pore enlargement. In comparison with the conventional devices, the proposed analysis system achieved better performance showing stronger correlation with the clinical grading. CONCLUSION: The proposed system is highly accurate and reliable for measuring the severity of skin pore enlargement. It can be suitably used for objective assessment of the pore tightening treatments.