High resolution passive facial performance capture

Lightweight wrinkle synthesis for 3D facial modeling and animation

“…4 illustrates a comparison between our results and the passive facial capture method [5]. Although the reconstructed wrinkle details from [5] are smoother than our synthesized wrinkle geometry, our synthesis results are visually very similar and provide high-quality clues of the input user expressions. Testing on Kinect data: We test our method on various captured RGB-D images and videos from different people to demonstrate the ability of our system to produce plausible 3D facial performances with detailed wrinkles.…”

“…Fig. 4 illustrates a comparison between our results and the passive facial capture method [5]. Although the reconstructed wrinkle details from [5] are smoother than our synthesized wrinkle geometry, our synthesis results are visually very similar and provide high-quality clues of the input user expressions.…”

“…Such approaches usually rely on multi-view stereo reconstruction with dense correspondences to recover the detailed wrinkles [3][4][5]. A lightweight solution, binocular stereo reconstruction, for marker-less facial animation is proposed in [2,9].…”

“…Laser scanning, multi-view stereo reconstruction and sophisticated deformation methods have been applied to the modeling of highly-detailed 3D facial models with fine-scale wrinkles [3][4][5]. However, these systems usually run in controlled laboratory environments by professional users with high-cost devices.…”

“…(a) The original captured images. (b) The results from[5]. (c) Wrinkle details generated by our method.…”