Sound Recognition becomes an important tool for intrusion detection or for\nthe monitoring of public premises exposed to personal hostility. It could further\nidentify different sounds. The main idea of the sound recognition\nprocess in a security system is to store samples of different sound signals in\nthe memory of the computer as references, and to be analyzed with respect to\ntheir frequencies components. In this paper, the sound signal of an unknown\nsource would be analyzed and compared with all the available reference samples,\nand then recognition is made according to the closest sample. The developed\nsecurity system consists of two main parts: the spectrum analyzer\nthat converts the sound signal to spectrograms. It is designed based on the\nreal-time analyzes, and the recognizer which compares the spectrograms and\ngives the decision of the recognition by using a special criterion. Experimental\nresults prove that the accuracy of the proposed system can be 98.33% for\nthe selected sample of signals.
Loading....