The goal of this project is to create a sophisticated Deep Learning system that can quickly and accurately evaluate Left Ventricle Ejection Fraction (LVEF), enhancing the diagnosis and management of cardiovascular diseases.Precise LVEF assessment is vital for forecasting the prognosis in conditions such as congestive heart failure, yet existing techniques are often slow and lack precision.
you can use Kaggle Notebook directly from HERE .
You will need to request access to the
EchoNet-Dynamic
dataset from Stanford University.
Once you have access to the data, download it and write the path of the "EchoNet-Dynamic" folder in the _dataRootPath variable in
Paths.
EchoNet-Dynamic Dataset
├── FileList.csv
├── VolumeTracings.csv
└── Videos
├── 0X1A0A263B22CCD966.avi
├── 0X1A2A76BDB5B98BED.avi
├── 0X1A2C60147AF9FDAE.avi
└── etc.
- train this model from Main and choose one model to train each time.
- HyperModel:
- Detecting (ES-ED) Frames
- U-Net Model:
- Detecting Left Ventricle
-
Add the Model Pathts in Paths
You can find All Models here:
- HyperModel Download best.pt
- U-Net ED
- U-Net ES -
Run API to open the local server using FastAPI.
-
Run the GUI file using flutter.
- Abanoub Gamal
- Kerolos Nabil
- Kerolos Helal
- Kerolos Waheed
- Yassa Kamille
- Ganna Muhammed
- Dr. Manal Mohsen Tantawi
- T.A. Radwa Reda Hossieny
- Secured Second Place in the AI in the Medical Field competition at Marathon Benha University 🥈.
- Presented and published our scientific paper at the 8th International Undergraduate Research Conference (IUGRC) at the Military Technical College. Also, We participated in the Military Technical College Science Exhibition, presenting both the paper and the project in the presence of the Minister of Defense. Additionally, We Were Invited to present and test the project at the Military Technical Hospital.
