[ICCV 2023]. Supervised Homography Learning with Realistic Dataset Generation. [Paper].
pip install -r requirements.txt
Please refer to Content-Aware Unsupervised Deep Homography Estimation (CAHomo).
- Original video data download links: [GoogleDriver], [BaiduYun] (key:gvor)
- Unzip the data and Run "video2img.py" to save the images to the directory "./Homo_data/img"
Be sure to scale the image to (640, 360) since the point coordinate system is based on the (640, 360).
e.g. img = cv2.imresize(img, (640, 360))
- Using the images in "train.txt" and "test.txt" for training and evaluation.
- The manually labeled evaluation files can be download from [GoogleDriver], [BaiduYun] (key:mrzz)
- Download links: [GoogleDriver], [BaiduYun] (key:j1zw)
- Unzip the masks to the directory "./Homo_data/mask"
| model | RE | LT | LL | SF | LF | Avg | Model |
|---|---|---|---|---|---|---|---|
| Pre-trained | 0.22 | 0.35 | 0.44 | 0.42 | 0.29 | 0.34 | [Google] [Baidu](key:qqed) |
You need to modify dataset/data_loader.py slightly for your environment, and then
python train.py --model_dir experiments/Base/
python evaluate.py --model_dir experiments/Base/ --restore_file Iter2_0.3445.pth
- Download links: [GoogleDriver], [BaiduYun] (key:gwry)
If you use this code or ideas from the paper for your research, please cite our paper:
@InProceedings{Jiang_2023_ICCV,
author = {Jiang, Hai and Li, Haipeng and Han, Songchen and Fan, Haoqiang and Zeng, Bing and Liu, Shuaicheng},
title = {Supervised Homography Learning with Realistic Dataset Generation},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
month = {October},
year = {2023},
pages = {9806-9815}
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent