InstRecognition Tool

An instrument recognition tool

• Note: More details on dataset, concept and implementation are available in the Jupyter notebook file IRT.ipynb.

Instructions and prerequisites

Prerequisites

• Anaconda 5.2.0 for Python 3.6 (py36_3)
• conda 4.6.3 (py36_0)
• Jupyter notebook 5.5.0
• numpy 1.14.3
• scikit-learn 0.19.1
• glob2 0.6
• librosa 0.6.3 from conda-forge
• IRMAS dataset from https://www.upf.edu/web/mtg/irmas

Additionally, it is required that the extracted IRMAS datasets are placed in the following folder:

• D:\College\Soft Computing\Data

The folder structure of Data folder should then look like:

• \ IRMAS-TestingData-Part1 \ Part1 \ {.wav & .txt files}
• \ IRMAS-TestingData-Part2 \ IRTestingData-Part2 \ {.wav & .txt files}
• \ IRMAS-TestingData-Part3 \ Part3 \ {.wav & .txt files}
• \ IRMAS-TrainingData \ {instrument label} \ {.wav files}
• \ DEFENSE \ {.wav & .txt files}

Instructions

1. Clone the project repository.
2. Open the IRT.ipynb file in a Jupyter Notebook.
3. From the toolbar, click on Cell, then on Run All.
4. Enjoy the awesomeness that is this project.
5. Grade generously.