Indian Folk Dance Form Classifer
CNN Model for classification of folk dances
This project is in reference to the Hacker Earth Deep Learning Challenge Identify the dance form
The classifier uses cnn to identify the image of a dance form.
Following are the class categories for identification:
- Manipuri
- Bharatanatyam
- Odissi
- Kathakali
- Kathak
- Sattriya
- Kuchipudi
- Mohiniyattam
The dataset for the same is available at: Data
Make sure your python enviroment fullfils the following deps. If not, install the following using 'pip install ___dep___'
Python Dependencies:
- Open Cv
- Pandas
- Tensorflow
- Numpy
After downloading the Dataset, place it in a folder named: data and run preprocess.py. When the preprocessing is finnished, you should have two new files with the name: features.py and labels.npy. Now run the model.py which will save the model to cnn.model dir. You are now ready to make predictions. To test an image, place it in the data/test/ folder and run predict.py.. Prediction results are saved in predictions.csv file