HMM and K-NN based Automatic Musical Instrument Recognition

• Published in: 2018 2nd International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC), 2018 2nd International Conference on pp. 350 - 355
• Main Authors: Jeyalakshmi, C., Murugeshwari, B., Karthick, M.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2018
• Subjects: Conferences; Feature extraction; Hidden Markov Model (HMM); Hidden Markov models; Instruments; K-Nearest Neighbor(K-NN); Mel frequency cepstral coefficient; Mel frequency cepstral coefficient (MFCC); Music; Musical instrument recognition system(MIS); Perceptual Linear Prediction (PLP); RASTA-PLP; Training
• DOI: 10.1109/I-SMAC.2018.8653725

Due to the tremendous increase in the amount of data, data indexing becomes a very challenging task at present. For data retrieval and indexing purpose, automatic musical instrument recognition plays a vital role. When the instrument is played, the computer has to recognize the sounds of the musical instrument. Systems constructed attempt to gather perceptually related information from musical instrument sounds and identify their sources. The goal of this paper is to separate four dissimilar instruments, explicitly flute, guitar, violin and piano. Since effective data management is a challenging one, robust feature extraction methods have to be utilized to achieve better performance. Feature extraction algorithms considered here are MFCC, PLP and RASTA-PLP and proposed system is analyzed using HMM and K-NN classifier. It is also evaluated for same and different musical note as dataset, recorded from 4 dissimilar instruments. The performance of the system is obtained according to accuracy and comparison done between HMM and K-NN methods.