Research And Implementation Of Facial Skin Quality Evaluation System Based On Deep Learning

At present,with the improvement of people’s living standards,people’s pursuit of beauty is getting higher and higher,especially for facial skin care.However,limited by the cognition of self-skin quality and medical knowledge,people often make wrong judgments and then buy wrong skin care products,which may even lead to the opposite skin care effect.Correct skin care needs to be based on accurate skin quality evaluation.How to overcome the inconvenience of traditional evaluation must rely on large-scale skin testing instruments in hospitals or beauty salons,and use advanced image processing and machine learning technology to develop a convenient,fast and higher Accurate small face skin quality evaluation system is a hot topic in the field of medical beauty.Based on the in-depth analysis of existing facial skin quality detection products and development technologies,the paper studies the human skin quality assessment method based on deep learning,focusing on the segmentation of skin quality detection targets and the level evaluation of targets in face images,and on this basis,a web-based assessment system is developed.In terms of theoretical research,the paper focuses on the segmentation of skin detection targets in face images and the level evaluation of targets.Among them,in view of the complex facial information and too many skin quality evaluation indicators,a neural network model often cannot meet the accuracy requirements of skin quality detection.The network is used to predict different skin quality indicators respectively,so as to achieve accurate evaluation of face skin quality indicators.Deeplabv3+ contains a large number of channel fusions in the network,which makes it more difficult to describe and extract features.Combined with the channel attention mechanism,an improved algorithm is proposed to further improve the accuracy and performance of segmentation.The VGG16 network with better learning ability and classification effect is used as the skin quality feature level classification algorithm,and each skin quality feature level is divided into 10 levels,and then sent to the skin quality evaluation model to comprehensively analyze the skin quality of the whole face.Evaluation.In terms of system design,the system functions of this paper include user login,user information upload,skin quality detection,result query and skin quality history record functions.Users can detect skin quality by uploading images.At the same time,the system records the results of each test,and summarizes these results into the historical record,which is convenient for users to check the changes in skin quality.At present,the software system has been basically designed.Tests show that the system can meet the requirements of online facial skin quality assessment for people anytime,anywhere.Compared with the existing face and skin quality detection equipment,the detection is more convenient,and the evaluation is more comprehensive and accurate.