4D capture - moving video
The 4D Vision project was initiated at the University of West of England in 2010 supported by HEFCE QR funding. The main goal of this project is to move existing 3D Photometric Stereo technology to the next stage and enable the capture of 3D faces in real time. In the near future, we expect to develop new imaging capabilities for high-speed and high resolution capture of facial movements, combined with robust multi-resolution analysis, realistic visualisation and fast interaction.
- Building a system for generic 3D facial macro- and micro-movement capture.
- Real-time reconstruction of moving shapes and 3D texture information at unprecedented sub-pixel resolution.
- Realistic rendering of moving faces with real-time visualisation and interaction.
- Create and maintain open access a research database showing specific interaction sequences of facial movements.
- Design and implement novel automated feature extraction and classification methods that exploit the spatio-temporal information of moving 3D faces and allow fast detection of macro- and micro-movements.
- Generate a new facial expression taxonomy using a robust and accurate model to map facial micro- and macro-movements to corresponding expressions.
Real-time capture and recovery
The capture and recovery of moving 3D faces in real-time involves the transfer and processing of high volumes of data. To address this challenge, we use a special combination of expertise both in hardware and software, where all reconstruction and analysis processing is run in parallel on a combined CPU and GPU processor. In addition to enabling high-performance recognition capabilities on moving 3D data, this approach has the advantage of using development platforms that will support portable applications.
Above: 4D rig in use with light sources indicated.
Facial Expression Modelling and Classification
Facial expression recognition is commonly undertaken within the 2D imaging domain. Recent developments at CMV , such as the Photoface and the 4D Vision projects, allow analysis of dense 3D surface information in both static and dynamic ways, respectively.
These systems employ a classification method which is both pose and illumination invariant, hence overcoming the limitations of 2D approaches. Unlike other commonly used 3D capture techniques, photometric stereo provides dense high frequency spatial information which allows the capture of fine details, such as wrinkles and transient furrows. This high density information also enables the extraction of curvature based features. Through statistical feature selection and SVM -based classification, we are then able to classify facial expressions with relatively high accuracy.