| Summary: | In this project, we conducted research and testing for facial retouching to treat imperfections such as wrinkles, scars and hyperpigmentation. Various colour spaces have been considered to accurately generate the skin mask and effectively segment the skin areas from the non-skin ones. Face detection has also been employed to facilitate working on the face specifically, without affecting the regions beyond it. To attain subtle and natural-looking results, different blurring filters were tested – Mean, Median and Gaussian Smoothing. The results obtained show that the proposed methods have the ability to enhance facial beautification. The structure of the report is as follows: Chapter 1 introduces the need for beautification computational techniques. Chapter 2 delves into the research behind the algorithms and techniques developed. Chapters 3-6 describe in detail the proposed facial retouching and background blur methods, the implementation, and results. Chapter 7 shows the GUI created to enable facial retouching. Lastly, Chapter 8 summarizes the key contributions of this project and potential works.
Keywords: Skin Segmentation, Image Processing, Blurring, Face Detection, MATLAB, Facial Retouching
|
|---|