|Description (include details on usage, files and paper references)||ANN: dynamic 2D/3D speaking face dataset with synchronized audio
We would like to announce a new facial biometric dataset that has:
* 1 second of 500 frame per second IR intensity video (600*600 pixels)
* 1 second of 500 frame per second registered depth images (600*600 pixels)
* synchronized 44.1 Khz audio
There are 77 participants looking at the camera.
Each is speaking the same passphrase (Ni hao).
There are 10 repetitions for each person (for a total of 770 sets of data).
There are an additional 26 (*10 repetitions) participants that were
moving their heads while speaking the same passphrase.
The data can be used for a variety of purposes: IR based facial recognition/verification
3D facial recognition/verification, speech-based recognition/verification,
facial motion analysis, etc.
If the data is used, please cite:
Jie Zhang, Robert B. Fisher; 3D visual passcode: Speech-driven 3D facial dynamics for behaviometrics, Signal Processing, 2019, 160: 164-177.