|
Project Information
|
The project is about the studying and developing of a personal indentifying application by facial recognition on mobile platform, Google Android for specific, accomplished by using available achievements and open resources in the computer vision field. The chosen recognition algorithm is the most advanced and promising one among many other, SURF - Speeded Up Robust Feature, enhanced with many improved minor algorithm and calculation methods for processing time reduction, and is combined with Haar Feature-base Cascade Classifier for multi-scale face detection. The result application is rather good and somewhat reliable, can be installed on any Android devices which must have at least 800Mhz CPU and 256Mb RAM. However there are still many issues to deal with before completion, such as reducing the processing time as it is taking too long for recognizing with large database, optimizing algorithms and improve using of hardware resources, in order to lower the minimum hardware requirement. And due to the limitation of OpenCV API for functioning Android device's camera, the application only support some specific models of handset, so far. |