Peer-Reviewed Publications & Open Datasets

The Research Behind FADA

FADA is built on a foundation of rigorous, peer-reviewed research. Explore the publications, methodologies, and datasets developed by our team at Hamad Bin Khalifa University to advance AI-driven prenatal diagnostics.

3
Publications
6,667+
Annotated Images
0.901
Best DSC Achieved
Open
Access Datasets
Publications

Research Papers & Datasets

Each contribution below underpins a component of the FADA pipeline — from dataset construction and image calibration to deep-learning segmentation models.

Paper 01
2024
Journal Article
Deep Learning Segmentation Prompt-based Learning Ultrasound

FetSAM: Advanced Segmentation Techniques for Fetal Head Biometrics in Ultrasound Imagery

Mahmood Alzubaidi, Uzair Shah, Marco Agus, Mowafa Househ

College of Science and Engineering, Hamad Bin Khalifa University, Doha, Qatar

IEEE Open Journal of Engineering in Medicine and Biology, Vol. 5 · 2024

FetSAM is a cutting-edge deep learning model aimed at revolutionizing fetal head ultrasound segmentation, elevating prenatal diagnostic precision. Trained on the largest fetal head metrics dataset to date, the model incorporates prompt-based learning and a dual-loss mechanism combining Weighted DiceLoss and Weighted Lovasz Loss optimized through AdamW, with class weight adjustments for better segmentation balance. Benchmarked against U-Net, DeepLabV3, and Segformer, FetSAM delivers unparalleled accuracy and sets a new benchmark for AI-enhanced prenatal ultrasound analysis.

Key Contributions
  • DSC of 0.90117 — outperforms U-Net, DeepLabV3, Segformer
  • Hausdorff Distance: 1.86484 · Average Surface Distance: 0.46645
  • Prompt-based learning + dual-loss optimization (Weighted Dice + Lovasz)
  • Targets fetal brain, CSP, and lateral ventricles (LV) segmentation
View PDF DOI Link
DOI: 10.1109/OJEMB.2024.3382487
Paper 02
2023
Conference Paper
Dataset Image Processing Calibration Fetal Biometrics

Conversion of Pixel to Millimeter in Ultrasound Images: A Methodological Approach and Dataset

Mahmood Alzubaidi, Uzair Shah, Hurmat Shah, Mowafa Househ

College of Science and Engineering, Hamad Bin Khalifa University, Doha, Qatar

IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) · 2023

Accurate fetal structure measurement hinges on precise pixel-to-millimeter (mm) conversion — an essential step in developing AI applications for automated fetal measurement. This paper presents a novel pixel intensity-based technique for converting pixel values to millimeters in fetal ultrasound images. Using a publicly available dataset, we add new labels providing pixel size in millimeters, producing an augmented dataset of 2,835 fetal head images — the largest and most diverse collection of its kind in the literature.

Key Contributions
  • 2,835 annotated fetal head ultrasound images
  • Mean pixel size of 0.144 mm (range 0.06 mm – 0.33 mm)
  • Novel intensity-based automated pixel-to-mm conversion
  • Foundation for AI-based fetal measurement in clinical settings
View PDF DOI Link
DOI: 10.1109/CIBCB56990.2023.10264909
Paper 03
2023
Dataset Article
Open Dataset Annotation CC BY 4.0 Zenodo

Large-scale annotation dataset for fetal head biometry in ultrasound images

Mahmood Alzubaidi, Marco Agus, Michel Makhlouf, Fatima Anver, Khalid Alyafei, Mowafa Househ

Hamad Bin Khalifa University · Sidra Medicine · University of Doha for Science and Technology

Data in Brief, Vol. 51, 109708 (Elsevier) · 2023

This dataset features 3,832 high-resolution ultrasound images (959×661 pixels) of fetal heads, focused on the brain, cavum septum pellucidum (CSP), and lateral ventricles (LV). It is provided in 11 universally accepted formats — including Cityscapes, YOLO, CVAT, Datumaro, COCO, TFRecord, PASCAL, LabelMe, Segmentation mask, OpenImage, and ICDAR — making it broadly compatible with medical imaging software and deep learning frameworks. Annotations were validated through a two-step process involving a Senior Attending Physician and a Radiologic Technologist.

Key Contributions
  • 3,832 high-resolution fetal head ultrasound images
  • Available in 11 standard CV / ML formats
  • Two-step expert validation (ICC: 0.859 / 0.889)
  • Open access on Zenodo under CC BY 4.0
View PDF DOI Link
DOI: 10.1016/j.dib.2023.109708
From Paper to Platform

How the Research Powers FADA

Each publication directly translates into a component of the FADA clinical pipeline.

01

Curated & Annotated Data

Our open-access dataset of 3,832 expert-validated fetal head ultrasound images forms the training foundation for every model in FADA.

02

Calibrated Measurements

Pixel-to-millimeter conversion enables FADA to output clinically meaningful measurements directly from raw 2D scans — no manual calibration required.

03

State-of-the-Art Segmentation

FetSAM powers FADA's anomaly detection — segmenting fetal brain, CSP, and lateral ventricles with industry-leading accuracy (DSC 0.901).

For Researchers

Citing FADA Research

If you use our methods or datasets in your work, please cite the corresponding publication.

FetSAM: Advanced Segmentation Techniques for Fetal Head Biometrics in Ultrasound Imagery 
Mahmood Alzubaidi, Uzair Shah, Marco Agus, Mowafa Househ. "FetSAM: Advanced Segmentation Techniques for Fetal Head Biometrics in Ultrasound Imagery." IEEE Open Journal of Engineering in Medicine and Biology, Vol. 5, 2024. doi: 10.1109/OJEMB.2024.3382487.
Conversion of Pixel to Millimeter in Ultrasound Images: A Methodological Approach and Dataset 
Mahmood Alzubaidi, Uzair Shah, Hurmat Shah, Mowafa Househ. "Conversion of Pixel to Millimeter in Ultrasound Images: A Methodological Approach and Dataset." IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB), 2023. doi: 10.1109/CIBCB56990.2023.10264909.
Large-scale annotation dataset for fetal head biometry in ultrasound images 
Mahmood Alzubaidi, Marco Agus, Michel Makhlouf, Fatima Anver, Khalid Alyafei, Mowafa Househ. "Large-scale annotation dataset for fetal head biometry in ultrasound images." Data in Brief, Vol. 51, 109708 (Elsevier), 2023. doi: 10.1016/j.dib.2023.109708.

From Research to Real-World Impact

Experience how peer-reviewed AI research translates into a working clinical tool for prenatal diagnostics.

Try the Platform Back to Home