Identification of sound source in machine vision sequences using audio information

Identifying the sound source location in the acquired image for further processing is of concern in image/video processing especially when tracking objects of vibrating or sound making nature in the scene becomes necessary. On the other hand relying on the picture content to identify such objects is not an easy task especially when there are similar objects in the acquired image. In this paper, we present a technique for identifying the location of sound source using beam-forming techniques by means of an array of microphones based on the processing of audio from the source itself. Upon estimation of the location in 3-D space, a projection technique is used to accurately locate the position of sound source in 2-D image frame coordination.