Ship Classification using Image Processing
This is a multi-class classification problem with 5 types of ships.
It was a 2-week nation-wide competition
- Applied SOTA practices like Transfer Learning, OpenCV, MixUp etc. on an image classification problem with 5 classes.
- Used architectures like ResNets, NASNets, DenseNets etc
- Used heavy Data Augmentations like Cutout, Colour Dimension shifts etc.
- Generated an ensemble of all the best performing models
https://www.kaggle.com/sandeeppat/ship-classification-top-3-5-kernel