Readme.md

BoundaryNet -- Baselines

CurveGCN

The code is tested with

• Python (3.5.x)
• PyTorch (1.0.0)
• CUDA (10.2)

cd baselines/CureveGCN/

Please install dependencies by

pip install -r requirements.txt

Usage

Initial Setup:

• Download the Indiscapes dataset - [Dataset Link]
• Place the
  • Dataset Images under doc_images directory
  • Pretrained Model weights in the checkpoints directory
  • JSON annotation data in data_splits directory

Training & Inference

python3 combined_train.py --exp train_experiment.json

• Any required parameter changes can be performed in the experiment file.

PolyRNN ++

The code is tested with

• Python (3.5.x)
• PyTorch (0.4.0)
• CUDA (10.2)

cd baselines/polyrnn-pp-pytorch/code/

Please install dependencies by

pip install -r requirements.txt

Usage

Initial Setup:

• Download the Indiscapes dataset - [Dataset Link]
• Place the
  • Dataset Images under doc_images directory
  • Pretrained CurveGCN/DACN Model weights in the checkpoints directory
  • JSON annotation data in data_splits directory

Training & Inference

python3 Scripts/train/train_ce.py --exp Experiments/mle.json

• Any required parameter changes can be performed in the experiment file.

DACN

The code is tested with

• Python (3.5.x)
• PyTorch (1.0.0)
• CUDA (10.2)

cd baselines/DACN/

Please install dependencies by

pip install -r requirements.txt

Usage

Initial Setup:

• Download the Indiscapes dataset - [Dataset Link]
• Place the
  • Dataset Images under doc_images directory
  • Pretrained Model weights in the checkpoints directory
  • JSON annotation data in data_splits directory

Training & Inference

python3 combined_train.py --exp train_experiment.json

• Any required parameter changes can be performed in the experiment file.