Musical Instrument Information retrieval using Neural Network

In this paper musical instrument recognition and retrieval for fifteen musical instruments from different instrument families are discussed. The system is implementing in three stages; first stage is pre-processing, second is feature extraction and third is recognition and retrieval. Musical instruments are retrieved using most important and distinguishable features like temporal and cepstral features. Kohenon self organizing map has been used as classifiers. The average accuracy is achieved for fifteen instruments are recorded 92.98%. The experimental results also show that the better recognition rate is obtained for LPCC as compared to MFCC and temporal for all the musical instruments.