Based on the predicted result of VGG16 model and reordering probably similar images via pre-trained Xception or ResNet model.
Can be downloaded and be used to construct model and labels as model_path and labelbin_path respectively in modules/get_probable_images_vgg16.py
version: [003_0320]
https://drive.google.com/drive/folders/13N8drxGjsNJ_8DeiGHKGihbMvC2ADmOM?usp=sharing