Design And Implementation Of Skin Beautifying Model Based On Skin Segmentation And Skin Quality Evaluation

With the development of economy and the improvement of living standards,the scale of the photography industry has a sustainable growth.Portrait photography mainly includes wedding photography,children’s photography,family photos,etc.The growth trend of market demand is obvious.The post-processing of portrait photography is mainly aimed at the beautification of human face and human skin.Large scale Studio Photos and personal photos need to be dealt with professional,fast and parameter adaptive automatic skin beautifying algorithm.There are several disadvantages of commercial or open source mapping algorithms in the market.Excessive beauty leads to skin distortion and unreserved texture.Manual intervention is required to adjust parameters,which cannot realize automation.The algorithms are unable to beautify the whole body skin area.To solve the above problems,the main contents of this paper are as follows:1)The key steps of skin beautification model with automatic parameter adjustment are designed,including face region of interest extraction,face skin quality evaluation,human skin segmentation and whole body skin beautification.The algorithms are implemented and evaluated.2)A skin segmentation algorithm based on u-net++ and subject location is designed and implemented.The human body region is segmented first,and then the skin is segmented in the human body region.The accuracy and efficiency of the algorithm are higher than other comparison methods.3)Based on the face key point alignment algorithm,four regions of interest are extracted: forehead,left cheek,right cheek and chin.The facial skin is obtained by calculating the intersection of the face region of interest and the human skin region.Four levels are respectively defined for the four indicators of human face skin,including skin color,wrinkles,pores and glossiness.The data set is trained and tested based on RESNET,and then the skin classification is calculated according to the weight of the four regions of interest,and the face skin evaluation model is constructed.4)According to the evaluation indexes of skin type,a skin beautification model was established.The key skin-grinding operator,degreasing operator and whitening operator in the skin beautification algorithm are optimized.According to the classification results of face skin type,the face and whole body skin are beautified by adaptive matching parameters,and then the output image is classified again to verify whether the algorithm beautifies the second,third and fourth level skin to the first level skin,and the effective action rate of the model is defined and calculatedThe experimental results show that the average accuracy of skin segmentation algorithm based on subject recognition is higher than 96%,the average accuracy of skin classification algorithm based on region of interest is higher than 88%,and the effective rate of parameter adaptive skin beauty model based on skin evaluation and skin segmentation is higher than 95%.The professional draughtsman scores the test set before and after skin beautification,and the results also show the effectiveness of the algorithm in this paper.The algorithm model proposed and implemented in this paper has reference significance and application value for related research and industry application.