Faceswap blending and quality improvement

Hi there, we have software with a trained model which do face swap on images. We need additional visual improvements. Blending improvement: The currently used blending method which is blending the faces is not good. Face unexpectedly shaking, moving from its place (small vibrations), bleary Etc… Your task is to fix the blending and improve the result image quality Check the added example video. The app is written in Python and your task is to fix or extend the existing code base. What are the required skills for the work? Git and Github Python 3.8 Numpy, PIL, PyTorch, OpenCV How can you reach the code? There is a private Github repository, where you will get access to. So you must have a Github account. We need your account’s name or e-mail address to give access to you. How to work? You need to create a new branch and work on it. When you are ready, raise a Pull Request on Github to the master branch and we will review your code. If it’s okay, we will merge it and you are done. How to test it? You will get the code without the trained model. You can test your code on our computer, via any desk. On our computer the git repository will be already configured, so you have to checkout your branch and pull your commits, finally run the application. Unfortunately we won’t provide you the model files. Important! We need someone who is free to work on the project. Please also note due to very strict contracts we can allow you to test your work via team viewer, or any desk remotely. Please also note that we can not pay for half or incomplete job. We are gladly pay if we are satisfied with your result, but we can not pay for 'I know it is not working but I worked so hard and so much'.