NON-RD_DCNN: A Deep Learning Model for Multi-Chord Music Classification

Abstract

A chord classification is a conventional category that designates some pieces of music as belonging to a shared tradition or set of conventions. It must be distinguished from musical style and form. There are numerous ways to categorize music into distinct genres. Pop, Hip- Hop, Rock, and Metal are some of the most popular chord genres. Ordering harmony files in view of their kind is a troublesome assignment with regards to music data recovery. Programmed harmony order is fundamental while endeavouring to separate music from a colossal assortment. It has pragmatic applications in an assortment of disciplines, for example, programmed labelling of an obscure piece of music (helpful for applications like Spotify, Wynk, and so on.).In this work, a novel model termed as NON-RD_DCNN is developed in two phases. In the first phase, the NONRD algorithm will be used for improvising the extracted data. The Further comes the second phase, in which the Deep-CNN model is used for predicting the output class label. For implementing the model, the GTZAN dataset as an input, which is the most popular music dataset that has file size of around 1 GB. The NON-RD_DCNN model gives the results like accuracy, loss, val_accuracy, val_loss.