This code has been completely written from scratch using PyTorch library. A sample of dataset has been already uploaded in folder of sample_dataset

1. Install requirements.txt by running

pip3 install -r requirements.txt

2. Follow the Jupyter notebook denoising_ae.ipynb for complete code
3. For loading dataset of origami change the variable origami_dataset_dir to the directory containing folders of input and output.
4. Similarly change the variable inp and out to the directory containing various images containing Network's input and Output images repsectively.

origami_dataset_dir = "MarowDataset"
inp='Input'
out='Output'

5. Dataset Directory info

    MarowDataset
    │
    ├── Input
    │   ├── img100_1.png
    │   ├── img100_2.png
    │   ├── ....
    |
    ├── Output
    │   ├── img100_1.png
    │   ├── img100_2.png
    │   ├──  ....

5. There are various helper functions written by me to aid the code. These can be found in file DAE_dataset_helper.py and DAE_model.py

## The classes imported below are used for dataloader, transformation and model

from DAE_dataset_helper import OrigamiDatasetGenerate
from DAE_dataset_helper import ToTensor,Resize, RandomBackground
from DAE_model import AugmentedAutoencoder

• Rohit Lal - WEBSITE
• Ruphan S