Feature visualization technique to approximate the best suitable deep model (Inception-V3, ResNet-50, and VGG-16) for a given data set. Instead of training all the deep models at you hand and compare results at the end, this method can be use to choose the best suitable model for your data set without training all models.
Please download the tf checkpoints at https://www.dropbox.com/s/6quqjrrwxizx9g6/save.rar?dl=0
For more details please refer to our paper https://github.com/LahiruJayasinghe/publications/blob/master/Deep%20Feature%20Learning%20for%20Cleanliness%20Classification%20in%20Restrooms.pdf