source code for paper https://link.springer.com/article/10.1007/s00521-018-3468-3

Deep transfer learning for military object recognition under small training set condition

training data in image200d, testing data in sample_images

output_graph_new.pb is model file which is get only by training model with fully conneted layer 200 steps. output_labels_new.txt record the labels in our paper

our code is default for training fully conected layer, if you want train more layer, you can replace retrain.py by that in other layer,and then

CUDA_VISIBLE_DEVICES=0 python retrain.py

we make a sample visual interface for testing

you can only

python ./open.py

and you can see

then click open image, chose an image for test

last,click test,you can get predict result

​
​