The Research And Design Of Makeup Digital Assistant System

In modern society,make-up has become the trend of the society.Women make up everyday and choose different make-up in different occasions.However,choosing the right make-up style is a difficult and time-consuming task,because it needs to put a set of cosmetics on a face to achieve it completely and then watch the effect.It takes a long time to get this done.At the same time,make-up is a skill that requires constant experimentation and practice to sum up experience,which in turn can gradually improve one's makeup level.According to the research on the market,the functions of the electronic products on the market,which can help women's make-up,are mainly focused on filling light to simulate different scenes,magnifying local details and virtual make-up.These products can help users draw better makeup to a certain extent,but the users of such products still need to rely on their own to summarize the experience of each make-up.Therefore,the subject of this project is to design a digital assistant system that can provide guidance when users make makeup,help users record the makeup process,make-up effects and evaluation,and help users accumulate experience and deficiencies in each make-up.So they can improve their own skill in each make-up.Based on the functional requirements of the system,this paper studies and implements the related algorithms in three aspects: facial spot search,eyebrow contour symmetry detection and digital makeup technology.1.The statistics and summary of the image features of each color channel in multi-color space with facial speckles were carried out.It was found that the brightness of normal skin under a* channel of CIE La*b* color space image was more uniform and the its brightness was lower than the speckles area.Due to the similarities between the face freckles image under the a* channel and the infrared image needed to find out the small target,the mean-contrast filter is creatively used to locate the facial blot,so as to suppress the skin area and enhance the speckle area,and then find the speckle target.According to the image features summarized,the LBP feature describing the local texture of the image combined with the SVM classifier was used to eliminate the false targets.Finally,a good facial spot search result is obtained and compared to the existing methods.2.According to Zhang Chuanfeng et al.proposed tht in eyebrow images without occlusion,there is a regularity between the grayscale distribution and the minimum grayscale of the eyebrow area,we completed the outline extraction of eyebrows.Then,the extracted outline is removed by remnant,and the minimal bounding rectangle image is extracted to obtain the pure eyebrow outline image.Based on the position symmetry and contour similarity inherent in eyebrows,the eyebrow outline image on one side is turned over,and the overlap area on the other side is compared to achieve the eyebrow contour symmetry post-detection in line with the expectation of the human eye.Test results were statistics.3.Digital make-up technology can pass the facial makeup effect in the template picture to the target picture so that the user can preview the effect of using the kind of makeup in advance to judge whether the makeup is suitable for herself.According to the method proposed by Zhen Beibei et al.,the template picture with makeup and the target picture to be painted are layered under the CIE La*b* color space to extract the structure layer,the detail layer and the color layer.Then synthesize each layer image by means of weight addition,color fusion.This method complete the transfer of makeup effects very well.Finally,this paper combines the voice dictation API of i Fly open platform,the semantic comprehension API of Diting robot and the voice synthesis API of i Fly open platform to realize the voice interactive function with makeup knowledge quiz.Then by applying the above research and implementation algorithm to the actual,this paper designed and implemented a makeup digital assistant system with makeup analysis,makeup learning summary,makeup effect prediction,voice interaction and so on.The system design concept is more novel,which can help users record and accumulate experience with their own make-up,so as to continuously improve make-up skill.The system is an innovation in cosmetic digital products.