PyTorch-Deep-Steganography-Using-Fourier-Transform

Usage

main.py [-h] [-use | -train] [-encode | -decode] [--datapath]
              [--num_train] [--num_val] [--size] [--bs] [--epochs] [--model]
              [--fourierSeed]

Optional Args

  -h, --help      show this help message and exit
  -use            Use for Inference
  -train          Train a new model
  -encode         Use encoder to encode secret image in cover image
  -decode         Use decoder to decode secret image from cover image
  --datapath      Path to Dataset folder
  --num_train     Number of training pairs to be created
  --num_val       Number of validation pairs to be created
  --size          Size of the images
  --bs            Batch Size
  --epochs        Number of Epochs
  --model         Path for the model file if you want to finetune
  --fourierSeed   Seed for generating the pseudorandom matrix for changing
                  phase in fourier domain

Tiny Deviations

• deviation between cover and contianer

  cover image	container image

• deviation between secret and revealed secret

  secret image	revealed secret image