Classification of Musical Instruments’ Sound using kNN and CNN

In the field of machine learning, the classification of musical instrument sounds is becoming popular. The aim of this work is to automate the identification and categorization of diverse musical instruments based on their acoustic characteristics. This classification approach entails gathering extensive audio data and employing various machine learning algorithms- kNN (k-Nearest Neighbor) and CNN (Convolutional Neural Network) to extract pertinent features. The proposed models can effectively distinguish between different musical instruments. These classification models find applications in various areas including music information retrieval, recognizing instruments in music production, and analyzing music performance. The kNN model achieves accuracy of 96% and CNN model achieves an accuracy of 88%. The F1-score of kNN and CNN model is 0.96 and 0.93 respectively.