In this project, I put together the MLS image deformation method and mediapipe's face detection module to perform some basic face image deformation. Only the first 20 landmarks were used in the demo due to time constraints. If the photodeform module is further sped up by the use of GPU computation such as pytorch, It should be possible to do deformations with more points.