Name		Name	Last commit message	Last commit date
parent directory ..
README.md		README.md
data_generation_util.py		data_generation_util.py
matching_data_generation.py		matching_data_generation.py
matching_data_visualization.py		matching_data_visualization.py
png_to_text.py		png_to_text.py
vocabulary_generation.py		vocabulary_generation.py

README.md

MATCHING Dataset Preparation

Please follow these instructions for data preparation for the Instance Matching Module.

Download the SketchyScene dataset following its instruction here. And you can place it anywhere you want.
We have provided the data in ~/Instance_Matching/data (the three json files). If you want to generate them by yourself, try:
```
python3 matching_data_generation.py --dataset 'all' --data_base_dir 'path/to/SketchyScene-dataset'
```
- You can set --dataset to be train/val/test.
For ground-truth data visualization:
```
python3 matching_data_visualization.py --dataset 'val' --data_base_dir 'path/to/SketchyScene-dataset'
```
- Set --image_id to the image you want. Otherwise, a randomly selected image will be displayed.