This repo is the source code of my research regarding disease detection from oil palm images. Try out the prediction at oil-palm-stem-prediction.fly.dev.
The prediction is performed using a machine learning model, SVM, with linear kernel. The feature used by the model is extracted from images using the ResNet50 pre-trained model. Unfortunately, I can't give the dataset used for training the model as it is private data. But you can use your own dataset to train the model. Put the dataset in the dataset/selected_dataset folder. Inside it, make sure it has both infected and normal folder.
Check how the feature extraction is performed in deep_feature_extractor.ipynb notebook. The model training is performed in skripsi.ipynb notebook.