Apple leaf disease - Keras CNN and Plant Pathology

Apple leaf diseases can be a large problem for apple orchards because the diseased plants can have direct repercussions on the number of apples that can be produced and the quality of those apples. Currrently, apple leaf diseases are discovered by manual labor which can take considerable amounts of time that can be used for production and raise labor costs. A deep learning model using Keras will be developed and trained on part of the training data and subsequently tested on a smaller part of the training data as the Kaggle test set is has only three images. The Keras model will attempt to classify a given leaf image from the newly created test dataset to a particular disease category. While this is a valiant effort, accuracy must be increased to satisfactory levels and this process will be performed with transfer learning using VGG19, ResNet152V2, Xception and InceptionResNetV2 in Keras.