Paper: CVPR2018, Learning Rich Features for Image Manipulation Detection
Code based on Faster-RCNN

This is a rough implementation of the paper. Since I do not have a titan gpu, I made some modifications on the algorithm, but you can easily change them back if you want the exact setting from the paper.

Python 3.6 TensorFlow 1.8.0

• Download vgg16 pre-trained weights from here
  • save to /data/imagenet_weights/vgg16.ckpt
• Two-stream neural network model: lib/nets/vgg16.py
  • noise stream's weights are randomly initialized
  • for accurate prediction, please pre-train noise stream's vgg weights on ImageNet and overwrite the trainable setting of noise stream after SRM conv layer
• Bounding boxes are predicted by both streams.
  • In the paper, RGB stream alone predicts bbox more accurately, so you may wanna change that as well (also defined in vgg16.py)
• Use main_create_training_set.py to create training set from PASCAL VOC dataset.
  • The generated dataset will follow the pascal voc style, which is also required by train.py
• Tensorboard file will be save at /default
• Weights will be save to /default/DIY_detaset/default

The code requires a large memory GPU. If you do not have a 6G+ GPU, please reduce the number of noise stream conv layers for training.

Dataset size: 10000, epoch: 3

I will update this repo a few weeks later after I installed the new GPU