Research Of Facial Skin Quality Assessment Methods Based On Intelligent Terminal

In recent years,China has become a global emerging cosmetics market.People's demand for beauty and skin care is also growing quickly.Traditionally,the judgment of their skin condition has a strong subjectivity,and lack of scientific knowledge about skin condition.Purchasing beauty products blindly may cause severe skin conditions.The skin quality index is the key to measure the condition of facial skin,it can provide the basis for purchasing beauty products.However,the current skin diagnostic instruments produced by domestic and foreign medical enterprises have the limitations of high cost,large volume,and low portability.The majority of home users can not use it conveniently.In view of the above situation,this paper designed an intelligent,portable human facial skin quality assessment system,the intelligent terminal can assess the skin condition,and provides skin care reference,achieving the function of intelligent facial skin quality assessment.This paper conducts a study on the evaluation methods of human facial skin quality.The Adaboost algorithm is first used to combine the Haar-like rectangle features to mark the face area and estimate the age of the face image.Then the three common skin parameters,oil,pigment,and roughness,are extracted and a machine learning algorithm is built.The skin quality assessment model analyzes the comprehensive performance of three different algorithms in training and testing,and uses the SVM algorithm to build the model finally.The Android terminal collects the user's face image through the front camera,then implements the function of face detection and age estimation through the system JNI layer.Finally,implementing the function of facial skin evaluation and returning the references to the user from the server.After testing,this skin quality assessment system can detect facial images stably.The accuracy rate of face detection reaches 96%,the accuracy rate of age estimation reaches 80%,and the accuracy rate of evaluation model is 81%.The results of assessment are objective.The functions of each module are normal in the intelligent terminal,and it can contact with the web server stably and efficiently,according with the real-time requirements.At the same time,the system is available compatibility,human-computer interaction interface is provided friendly,which provides the user with a convenient and efficient experience of the skin quality assessment.