An Official PyTorch Implementation of “EllipseNet: Anchor-Free Ellipse Detection for Automatic Cardiac Biometrics in Fetal Echocardiography (MICCAI-2021)”
by Jiancong Chen, Yingying Zhang, Jingyi Wang, Xiaoxue Zhou, Yihua He*, Tong Zhang*.
This repo contains the official pytorch implemetation for EllipseNet.
Please refer to https://git.openi.org.cn/OpenMedIA/EllipseFit.Mindspore for a MindSpore version. Please be noted that the MindSpore version is not an Ellipse Detection Framework but using a 2D Unet to train a segmentation network and then using ellipses to fit the segmentation results.
Framework
Installation
Please refer to INSTALL.md for installation instructions.
Experiments
Methods |
Setting |
DiceT |
DiceC |
Diceall |
Pavg |
EllipseNet (exp6) |
only IoU loss |
0.8813 |
0.8520 |
0.8666 |
0.8855 |
EllipseNet (exp1) |
w/o IoU loss |
0.9338 |
0.9108 |
0.9224 |
0.8841 |
EllipseNet (exp3) |
w/ IoU loss |
0.9430 |
0.9224 |
0.9336 |
0.8949 |
Training and Evaluation
Prepare the elliptical dataset in coco-format. An example script is given in scripts/prepare_label.ipynb. We provide scripts for all the experiments in the experiments folder.
Usage:
chmod +x experiments/miccai21/*.sh
./experiments/miccai21/exp3_base_theta5_iou.sh
Reproduction
If you need the docker for reproduction, please contact via email. We will provide the docker image.
License
EllipseNet itself is released under the MIT License (refer to the LICENSE file for details).
Portions of the code are borrowed from CenterNet, Rotated_IoU, human-pose-estimation.pytorch (image transform, resnet), CornerNet (hourglassnet, loss functions), dla (DLA network), DCNv2 (deformable convolutions), tf-faster-rcnn (Pascal VOC evaluation) and kitti_eval (KITTI dataset evaluation). Please refer to the original License of these projects (See NOTICE).
Citation
If you find this project useful for your research, please use the following BibTeX entry.
Chen J., Zhang Y., Wang J., Zhou X., He Y., Zhang T*. (2021) EllipseNet: Anchor-Free Ellipse Detection for Automatic Cardiac Biometrics in Fetal Echocardiography. In: de Bruijne M. et al. (eds) Medical Image Computing and Computer Assisted Intervention – MICCAI 2021. MICCAI 2021. Lecture Notes in Computer Science, vol 12907. Springer, Cham. https://doi.org/10.1007/978-3-030-87234-2_21
Contact
If you have any questions about this paper, welcome to email to zhangt02@pcl.ac.cn