Interactive automatic fetal head measurements from ultrasound images using multimedia computer technology

Area of Science:

• Medical imaging
• Fetal ultrasound technology
• Computational anatomy

Background:

• Accurate fetal head measurements are crucial for assessing gestational age and detecting potential growth abnormalities.
• Traditional manual measurement methods can be time-consuming and prone to inter-observer variability.
• Automated tools are needed to improve efficiency and consistency in prenatal diagnostics.

Purpose of the Study:

• To develop and evaluate an automated tool for detecting fetal skull boundaries in ultrasound images.
• To measure biparietal diameter (BPD) and head circumference (HC) using the developed algorithm.
• To assess the performance and speed of the algorithm on different hardware platforms.

Main Methods:

• Utilized active contour models for automatic detection of inner and outer fetal skull boundaries.
• Implemented and optimized the algorithm on a Sun SparcStation 20/71 and a MediaStation 5000 (MS5000).
• Compared automated measurements of BPD and HC with a gold standard and between the two hardware platforms.

Main Results:

• The algorithm achieved a significant reduction in execution time from 32 seconds to 248 milliseconds on the MS5000.
• Mean differences for BPD and HC measurements were 1.43% and 1.96%, respectively, compared to the gold standard.
• No significant differences were found in BPD and HC measurements between the Sun SparcStation and the MS5000.

Conclusions:

• The developed automated tool accurately measures fetal BPD and HC using active contour models.
• The substantial reduction in processing time makes the algorithm a practical tool for real-time ultrasound applications.
• This technology has the potential to enhance the efficiency and reliability of prenatal ultrasound examinations for sonographers.