An automatic facial skin defect detection system is proposed. The proposed method can automatically detect human faces in the input image. Facial features including eyebrows, eyes, nose and mouth are used to locate regions of interest (ROIs). Since specific facial skin defects are most likely to appear in a certain ROI, different features are used in different ROIs for facial spot and wrinkle detection. To detect facial spots effectively, their features, which are represented using color information, are extracted and then classi_ed using an SVM classi_er. To detect wrinkles correctly, an adaptive threshold is used on orientation-sensitive features. Experimental results demonstrate the effectiveness of the proposed approach.
|頁（從 - 到）||717-722|
|期刊||ICIC Express Letters, Part B: Applications|
|出版狀態||Published - 2011 六月|
All Science Journal Classification (ASJC) codes
- Computer Science(all)