Related Experiment Video
Updated: Sep 12, 2025
06:43
High Frequency Ultrasound for the Analysis of Fetal and Placental Development In Vivo
Published on: November 8, 2018
10.4K
Fada: Fetal Accurate Detection AI for Automated Ultrasound Image Analysis and Reporting
Uzair Shah1, Mahmood Alzubaidi1, Elyas Alamri1
1College of Science and Engineering, Hamad Bin Khalifa University, Doha, Qatar.
Studies in Health Technology and Informatics
|August 8, 2025
Summary
This study presents Fetal Accurate Detection AI (FADA), an AI framework that generates clinical descriptions from fetal ultrasound images. FADA demonstrates high accuracy, improving automated prenatal care reporting.
Area of Science:
- Medical Imaging and Artificial Intelligence
- Prenatal Diagnostics
Background:
- Fetal ultrasound interpretation requires specialized expertise.
- Automated reporting can enhance efficiency and consistency in prenatal care.
Purpose of the Study:
- To introduce Fetal Accurate Detection AI (FADA), an AI framework for generating clinical descriptions from fetal ultrasound images.
- To evaluate FADA's performance in diverse anatomical views and imaging modalities.
Main Methods:
- Utilized the Bootstrapping Language-Image Pre-training (BLIP) architecture, fine-tuned on 38,723 images.
- Incorporated Low-Rank Adaptation (LoRA) for image encoder parameter tuning.
- Enhanced the text decoder for medical vocabulary and structured reporting.
Main Results:
- Achieved a BLEU score of 0.984 on the validation set and 0.9589 on the test set.
- Demonstrated high efficacy in aligning AI-generated descriptions with expert annotations.
- Showcased successful application across trans-abdominal and trans-vaginal imaging.
Conclusions:
- FADA represents a significant advancement in automated medical reporting for prenatal care.
- The framework effectively bridges computational capabilities with clinical needs.
- Sets a new benchmark for AI-driven prenatal diagnostic tools.
