, Christian Wallraven Implementation of Motion Detectors

Combining 3D Scans and Motion Capture for Realistic Facial Animation

We present ongoing work on the development of new methods for highly realistic facial animation. One of the main contributions is the use of real-world, high-precision data for both the timing of the animation and the deformation of the face geometry. For animation, a set of morph shapes acquired through a 3D scanner is linearly morphed according to timing extracted from point tracking data rec...

The MPI VideoLab - A system for high quality synchronous recording of video and audio from multiple viewpoints

The MPI VideoLab is a custom built, flexible digital videoand audio recording studio that enables high quality, time synchronized recordings of human actions from multiple viewpoints. This technical report describes the requirements to the system in the context of our applications, its hardwareand software equipment and the special features of the recording setup. Important aspects of the hardw...

3D-Reconstruction of Faces: Combining Stereo with Class-Based Knowledge

The role of characteristic motion in object categorization.

We report three experiments where we investigated the role of movement in object recognition. Previous studies have suggested a distinct and separate mechanism for object motion encoding, related to the action or motor-based system. To date, however, the role of an object's motion in long-term memory representations has not been explicitly tested. Here we were specifically interested in whether...

Learning from humans: computational modeling of face recognition.

In this paper, we propose a computational architecture of face recognition based on evidence from cognitive research. Several recent psychophysical experiments have shown that humans process faces by a combination of configural and component information. Using an appearance-based implementation of this architecture based on low-level features and their spatial relations, we were able to model a...