Kaggle competition: Global Wheat Detection - identification of wheat heads from images
- Encoder-Decoder model implemented in tensorflow keras (2.2.0)
- Adam Optimizer: starting dynamic learning rate (ReduceLRonPlateau, starting rate = 0.01) and Early Stopping
- Test accuracy = 84%