Abstract

Emerging multi-modal signal processing applications require a sustained effort on the part of the developer to realize and deploy an application. A rapid prototyping platform will reduce the effort, cost, and time required to develop and deploy an application. In this paper, a rapid prototyping platform is developed for realizing a multi-modal signal processing application that involve real time interfacing of multi-modal signals both at the input and the output. The platform allows the designer to simulate various applications and produce the required product only after entire testing has been done. A portable intelligent meeting capture system that can be rapidly deployed in smart meeting rooms is implemented on this platform. The setup consists of a microphone array which computes the two-dimensional direction of arrival (DOA). The azimuth and the elevation angles are computed using advanced signal processing algorithms like GCC-PHAT, MUSIC which are implemented on the Real Time Operating System (RT-OS). The DOAs are communicated to a wireless networked camera which steers in real time towards the active speaker. Performance evaluation of the rapidly prototyped system is tested in real time meetings in terms of average error deviations in the DOA. The accuracy of the results indicate further miniaturization of the system. The possibilities of using this platform for developing multi-modal signal processing in general is also described.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call