Research On Face Deformation Privacy Protection Algorithm Based On Triangulation

In recent years,due to the increasing popularity of smartphones and other mobile devices,the widespread application of face recognition technology has brought about the era of brushing the face.Taking personal images and sharing them on social platforms has become a common part of our daily lives.However,the shared image contains personal bio-private information.The personal bio-information is different from the password that is usually set.If the password is leaked,the password can be changed immediately.If the bio-information is leaked,it will be a lifetime leak and the user's personal information will be safe.Putting in greater uncertainty leads to a series of privacy risks.Therefore,the privacy protection of the face has attracted more and more attention from researchers in recent years.The face privacy protection mainly hides or modifies the face features in the image through technical means,thereby protecting the face identity information in the image from being leaked.As people share images more frequently on social media,the issue of face privacy protection raises concerns that social media has the need to provide users with privacy protection tools.The main work of this paper is as follows:Firstly,the research status of existing face privacy protection methods is expounded,and its advantages and disadvantages are analyzed.Then the current status of face detection and face key point detection is introduced.The excellent characteristics of Delaunay triangulation for face deformation are analyzed.Finally,the method of face privacy protection based on image filtering,which is common in life such as pixelation and blurring,is discussed.The privacy protection effect and performance are evaluated,and the reasons of irreversible,unsafe and poor robustness are analyzed.Then,through the simulation analysis of face substitution and face synthesis,the triangulation-based face deformation privacy protection algorithm will generate a new privacy protection problem for face replacement;face synthesis will be in the non-face area during the synthesis process.The problem of shadow generation is generated.The face synthesis algorithm is improved.A face deformation privacy protection algorithm based on face region triangulation is designed.The algorithm is based on face frame setting,face key point detection and face area triangulation.Image processing methods such as fractionation,affine transformation,image erosion,etc.,and finally obtain a privacy-protected image.The experimental results show that the improved algorithm does not change the visual effect of the whole image while deforming the face,and the deformed face has low similarity and unrecognizable height distortion to the original face.Compared with the privacy protection method of image filtering,the security and robustness are better.Compared with the face synthesis algorithm,the non-face area shadow is solved under the premise of ensuring privacy protection performance,security and robustness.The problem is better visually and more in line with the privacy protection requirements.Finally,based on the improved triangulation face deformation privacy protection algorithm,a face privacy protection system suitable for social media is developed and implemented.According to the application scenario and user requirements,the data and the database were fully conceived,and the design of the system was completed.