In this demonstration, the power and flexibility of the CEVA-MM3101 is shown enabling an advanced face recognition capability, which could be implemented for video conferencing, gaming social networking or automotive applications. A simple HDMI connection between the video input device — a handheld video camera, for example — and the CEVA-based board is all that is required.
The high performance combination of the video processing engine and optimized software supports a wide range of face recognition features, including rotate, tilts, and head nodding. It also supports recognition of faces stored in a database, and is unlimited in the number of faces that can be detected. Images as small as 160 pixels can be recognized with the CEVA solution.
