Classification of Foliar Diseases in Apple Leaves

• Executed deep learning models using ResNet architecture and cross-entropy loss function to accurately classify apple foliar diseases from imagery, achieving a 96.97% accuracy through rigorous 5-fold cross validation

• Innovated a data augmentation strategy utilizing techniques including flipping, rotation, blurring, and adjustments to color/contrast to effectively address data insufficiency and data imbalance throughout project

• Engineered an advanced data augmentation pipeline levering the Albumentations library in Python, resulting in quintupling the size of the training dataset to significantly enhance the robustness of the deep learning model