3D estimation of facial muscle parameter from th 2D marker movement using neural network

摘要

Muscel based face image synthesis is one of the most realistic approach to realize life-like agent in computer. Facial muscel model is composed of facial tissue elements and muscels. In this model, forces are calculated effecting facial tissue element by contraction of each muscel strength, so the combination of each muscle parameter decide a specific facial expession. Now each muscel parameter is decided on trial and error procedure comparing the sample photograph and generated image using our Muscel-Edior to generate a specific face image.