- Data preprocessing and augmentation
- Dimention Reduction using Autoencoders
- Exploiting pretrained models
The goal was to classify the Graspevine leaves using the relevant dataset. We have used varisous Deep learning modesl for classification and data augmentation. The MobileNetV2 model got sufficient percision(90%). Also the state of the art for this dataset have used MobileNetV2 as the backbone in order to achieve best possible result.
Relevant pictures and statistics are shown in the notebook.