PyTorch implementation for reproducing Perspective Network (PFNet) results in the paper Rethinking Planar Homography Estimation Using Perspective Fields by Rui Zeng, Simon Denman, Sridha Sridharan, Clinton Fookes.
- Download the COCO 2014 dataset used in the paper.
- Download our trained weights from Baidu Cloud(Extraction code: enkv) and put it in the root directory
- Python 3.6
- Pytorch 1.8.0
- Train a PFNet model on the COCO dataset from scratch:
python train.py --dataset=/path/to/COCO
- Evaluate the model checkpoint
python evaluate.py --dataset=/path/to/COCO --model=./pfnet_0200.pth
MACE: 1.21
results |
---|