CAP6610-Project-Identify the style of song

• This project is to classy style of music which are Progressive and non-Progressive. First, the songs of music are transferred into images by Libros. Then, the classifiers such as CNN, GoogLeNet, and ResNet-50 try to identify the style of the song by the transferred images from songs.

Install Dependencies

• Python3
• librosa
• matplotlib
• pandas
• opencv-python
• pydub
• torch
• torchvision
• openpyxl
• tensorflow
• keras

Install Tools

• ffmpeg

  https://www.ffmpeg.org/

Setup

1. Clone repository

git clone https://github.com/ryonakennedy/CAP6610-Project

2. Change to working directory from clone

cd CAP6610-Project

3. Install project requirements (linux example)

pip3 install -r requirements.txt

apt-get install ffmpeg

4. Create folder for music files used to extract features from

mkdir Dataset

5. Copy music files into Dataset folder retaining music folder hierarchy (make sure only audio files are in directories or script may halt/fail)

Dataset
├── Not_Progressive_Rock
└── Progressive_Rock_Songs
Dataset/Not_Progressive_Rock/
├── Other_Songs
└── Top_Of_The_Pops

Method 1

Extracting Data for Classifiers

For extracting features from sound files

./ExtractFeatures.py

Running Classifiers

• CNN2_BCELoss

./CNN2_BCELoss.py

• Fisher Discriminant, Random Forests, Linear SVM, Regression Tree

./SciKitLearnClassifiers.py

Method 2

Extracting Data for Classifiers

Extract Raw Data

./save_data.py

Extract snippets from Raw Data

./save_snippets.py

Running Classifiers

• Classifier-Attention-Based CNN

Train

./train.py

Evaluate

./evaluate.py