Customer Reviews

Data-Driven 3D Facial Animation

5 star:		(0)
4 star:		(1)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

The book sits squarely at the intersection of two fields. Text to speech (TTS). And 3d animation, specifically of human faces. Both fields are still of course actively worked on, since neither is perfect. The difficulty here is to integrate the two. The papers in the book describe the many challenges facing researchers.

One of which is simply to do this in real time, or as close to real time as possible. A key motivation here is the rise of 3d worlds, like Second Life. Currently, all those have crude renderings of faces; and voice output from characters tends to be minimal and poorly done.

Going thru the book shows that implementing the subtle facial movements is tough. Because ultimately, evolution has hardwired us to be very perceptive of facial expressions. There are nuances that people can easily detect in actual faces, that are currently hard to express in code and generated images. The concept of blendshapes appears in many places in the text. Referring to building faces out of a rich eigen set of fundamental face images.