Speaker localization for microphone array-based ASR: the effects of accuracy on overlapping speech
Proceedings of the 8th international conference on Multimodal interfaces
Region of interest extraction and virtual camera control based on panoramic video capturing
IEEE Transactions on Multimedia
An introduction to biometric recognition
IEEE Transactions on Circuits and Systems for Video Technology
| Hi-index | 0.00 |
Our research work deals with the problem of automatic speaker tracking by camera. Such tracking systems do exist nowadays, but they suffer from a number of mechanical problems. To overcome these problems, we thought of employing a virtual tracking system using a fixed camera that does not require any mechanical part. But, is it possible to track a moving speaker with a fixed camera? If the task is already difficult enough with a mobile camera, how difficult would it be with a fixed camera? Trying to find a solution to the problem, we have proposed and conceived a virtual tracking system, which is able to ensure the required task by using only two cardioid microphones and a classic video camera. In this virtual tracking system, the task of speaker tracking is ensured by the orientation of the region of interest ROI of the camera towards the active speaker; we have called this method the virtual region of interest VROI based technique. Experiments show the good performance of the new virtual technique.