Skip to content

Latest commit

 

History

History
29 lines (23 loc) · 1.26 KB

README-TRAIN.md

File metadata and controls

29 lines (23 loc) · 1.26 KB

OMR Project Team 4

 We have used SVM in the classifier with linear kernel as it's fast and very accurate in our case.


1. How to train the classifier?

 1. You have to create a folder of the dataset containing folders for each symbol as each folder name is the label of the symbol.

  - You can find our naming convention here .

 2. Each image should be binarized and inverted (symbols = white, background = black).

 3. Then run features.py file as follow:

    python features.py [Dataset path]

2. The Dataset that we have used:

 - We have generated and collected our own dataset.

 - Our data set is a collection of printed and handwritten music notes.

 - You can find it here .


3. How much time does it take to train The classifier?

 - For a dataset with size about 48K images (41 MB), It takes about 5-6 minutes.


4. The Hardware that we used for training:

 - CPU: core i7-8750H

 - RAM: 16 GB