The Design And Implementation Of Portable Face Skin Detection And Evaluation System

The pursuit of beauty is the eternal topic of human, especially women. Face as the visual reflection of people’s look, is even more concerned; the skin quality not only affects the look, but also reflects the state of health. Face skin detector can give scientific skin care proposal based on the skin condition, so it is widely recognized and has broad market prospects. Currently, the face skin detector products on the domestic and foreign market are mainly divided into two categories, one is metal detecting skin detector based on bioelectrical impedance measurement, which has a cheaper price but single detection function and the accuracy is also low; the other one is large detector based on image processing, which has comprehensive function but it is complicated to operate, expensive and with a huge volume, often suitable for large beauty salons and medical institutions, and it does not apply to individual people.In view of this, this paper mainly launches the research from three aspects: the image processing algorithm of skin detection、the evaluation index of the skin model and application of model based on mobile terminal in order to obtain the detection and evaluation scheme of portable face skin with multiple index, high precision. And the designed reasonable evaluation model can provide initial reference for faces skin analysis of different age group and different regions. The main research work of this paper is as follows:(1)It designed face detection algorithm based on image processing of skin index. The detection of oil, pigment index is based on the model of S color space and V color space of HSV, which based on the logic operation; The testing pore size index are based on the combination of detection means witch are the dual threshold segmentation and morphological operation; The roughness index is based on the value analysis of feature on gray level co-occurrence matrix. And it takes the system of Taiwan CBS-1800 detector as criterion, in order to verify the effectiveness of the skin index of the detection algorithm in this paper.(2)Design a the evaluation model of human skin index. Firstly, detect oil, pigment, pores and the roughness of the 60 people’s facial skin image of 20-25 years old,set a sample data, then use the SVM classifier, BP classifier and KNN classifier for training and testing the samples, Comparative analysis the performance of three kinds of classifiers from classification accuracy and time complexity, the skin evaluation model which based on SVM classifier has optimal performance, the correct classification rate is 82.7308%.(3)Having researched Implementation Scheme of the portably detect skin type in human face and Evaluation System. To implement this system, dividing this system into 3 parts, including Image Collection Device, Host Computers, and Android Client. Firstly, human face image will be sent to the mobile phone app by image collection device through Bluetooth Module. Secondly, the face image will be sent to the Host Computers by the app, then Host Computers will handling this image and assess the result. Finally, download the specific index and result to app to display and realize the detection and analysis of skin type.