Ultrasonic marker pattern recognition and measurement using artificial neural network
SIP'10 Proceedings of the 9th WSEAS international conference on Signal processing
WSEAS Transactions on Information Science and Applications
Hi-index | 0.00 |
Measurement of nuchal translucency (NT) thickness in the first trimester of pregnancy has recently been proposed as the most useful marker in the early screening for fetal chromosomal abnormalities. Ultrasonic measurement of NT thickness is currently performed by manually tracing the two echogenic lines and locating the electronic calipers on the inner edges of these lines. The drawbacks of this method are inter- and intra-observer variability, and its inefficiency. In particular, accurate caliper placement requires highly skilled operators since the border of the nuchal translucency layer is very thin. We present a computerized method of detecting the border of the NT layer by minimizing a cost function using dynamic programming. Local measurements of intensity, edge strength and continuity are extracted and become weighted terms in a cost function. Our method can obtain accurate and reproducible results, and has been validated by computing the correlation coefficient between manual and automatic measurements.