Folder program-2020 is unavailable.
ATS-CNN's codes are in Folder ATS-CNN_SIQAD. (ATS-CNN: Asymmetric two-stream network for screen content image quality assessment based on region features)
getTFfile_SIQAD_unite_random.py converts image data into tfrecord files. All path values in this file need to be reconfigured carefully.
run getTFfile_SIQAD_unite_random.py
mytwo_unite_san_cha_random.py builds the CNN network framework and trains the network through the generated tfrecord file data. The path value in this file need to be reconfigured.
Also, SIQAD_ready_unite_random.py needs to be placed in the same folder as mytwo_unite_san_cha_random.py. Because SIQAD_ready_unite_random.py contains some parameters and functions necessary to train the model.
run SIQAD_ready_unite_random.py
mytwo_test_all.py is the test file. The data of the trained network model is required when the file is run. In addition, mytwo_unite_san_cha_random.py can become mytwo_test_all.py after some code changes.