This repository contains a Convolutional Neural Network (CNN) model for classifying different musical notes. The CNN model is trained on a dataset of music symbols, enabling automated recognition of note types in sheet music.
Integratring with a simple GUI, this software allows you to simply classify different notes and get a short tutorial on each type of note.
In the realm of music notation, the automated recognition of musical notes is a fascinating application of deep learning. This project aims to classify different music symbols representing notes using a CNN model. The trained model can provide insights into the types of notes present in sheet music.
The dataset used for training and validation comes from the Music Notes Datasets available on Kaggle. It contains five different music symbols:
- Whole Note
- Half Note
- Quarter Note
- Eight Note
- Sixteenth Note
Each category consists of 1000 data samples, with 800 for training and 200 for validation.
-
Clone this repository to your local machine:
git clone https://github.com/rambodazimi/music-note-classification.git
-
Navigate to the project directory:
cd music-note-classification
The GUI application allows you to upload an image containing a musical note, predicts its class, and provides a description of the note. Just run the Python file using the following command on your terminal:
python model.pyThe CNN model architecture used for music note classification is as follows:
Layer (type) Output Shape Param
conv2d (Conv2D) (None, 62, 62, 32) 896
...
dense_1 (Dense) (None, 5) 645
Total params: 1,240,869
Trainable params: 1,240,869
Non-trainable params: 0
After training for 10 epochs, the model achieved a validation accuracy of approximately 0.9, demonstrating its effectiveness in classifying musical notes.