This is an unofficial Pytorch implementation of View Adjustment Prediction Model in Camera View Adjustment Prediction for Improving Image Composition(2021).

View Adjustment Prediction Model, which we named as VAPNet, predicts whether an image needs adjustment, which adjustment, and how much adjustment is needed.

• Columns from Left to Counter-Clockwise means F1-Score

• GAICD [Paper][Github]
• FCDB [Paper][Github]
• Open Images[Homepage][CVDF]

git clone https://github.com/PROLCY/VAPNet-Pytorch.git
cd VAPNet-Pytorch && mkdir weight

pip install -r requirements.txt

Download pretrained model in the directory weight

python demo.py {image_dir_path}

You can check the inference result in terminal.

Predicted view adjustment is as follows.

If you are interested in this repository, please contact [email protected]