|
|
--- |
|
|
license: cc-by-nc-sa-4.0 |
|
|
language: |
|
|
- en |
|
|
- es |
|
|
pipeline_tag: object-detection |
|
|
tags: |
|
|
- aorticvalve |
|
|
- echocardiography |
|
|
- deeplearning |
|
|
- fasterrcnnresnet50 |
|
|
--- |
|
|
# 🫀 Aortic Valve Detection Model (AoVdetector) |
|
|
|
|
|
|
|
|
## Description |
|
|
|
|
|
This model detects the **aortic valve in echocardiogram images** from the following views: |
|
|
- PLAX (Parasternal Long Axis) |
|
|
- PSAX (Parasternal Short Axis) |
|
|
- 3CH (Three-Chamber View) |
|
|
|
|
|
The detection model is based on **Faster R-CNN with a ResNet-50 backbone**, pretrained on the **COCO dataset (COCO_V1)** and fine-tuned for cardiac ultrasound images. |
|
|
|
|
|
## Model Details |
|
|
|
|
|
- **Repository:** https://github.com/CardiovascularImagingVallHebron/AoVdetector |
|
|
- **Developed in:** Python |
|
|
- **Model type:** Object Detection |
|
|
- **Architecture:** Faster R-CNN (ResNet-50) |
|
|
|
|
|
## Sample results |
|
|
|
|
|
 |
|
|
|
|
|
## Model Performance |
|
|
|
|
|
- **Label Match Accuracy:** 92% |
|
|
- **View Type Accuracy:** 98% |
|
|
|
|
|
## Contact |
|
|
|
|
|
For any questions or inquiries, feel free to reach out to: |
|
|
**Pere Lopez Gutierrez** |
|
|
📧 pere.lopez@vhir.org |
|
|
|
