The dataset consists of 4771 images spanning 8 classes. The images are 64 * 64. Data augmentation techniques such as horizontal flips, blurring and variations in hue and saturation were applied. The images were scraped from different search engines like Bing, DuckGO and Flickr. (No copyright was intended).
| Class | Number of Images |
|---|---|
| Gond | 628 |
| Kalighat | 636 |
| Madhubani | 616 |
| Mural | 669 |
| Pattachitra | 630 |
| Phad | 504 |
| Tanjore | 554 |
| Warli | 534 |
| Total | 4771 |
The final report of the project is also available here.
ECE 6254 - Statistical Machine Learning
Advisor: Dr. Larry Heck
Final Project by Nimisha Pabbichetty, Ashwin Kulkarni, Siddhi Pandare and Aakash Venkataraman.
Georgia Institute of Technology