HarmonicEchoNet: Leveraging harmonic convolutions for automated standard plane detection in fetal heart ultrasound videos

Download PDF
Open Webpage

Md. Mostafa Kamal Sarker, Divyanshu Mishra, Mohammad Alsharid, Netzahualcoyotl Hernandez-Cruz, Rahul Ahuja, Olga Patey, Aris T. Papageorghiou, J. Alison Noble

Published: 01 Jan 2025, Last Modified: 08 Oct 2025Medical Image Anal. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•We introduce HarmonicEchoNet, a lightweight model for fetal heart standard view detection.•HarmonicEchoNet integrates harmonic convolution blocks and hscSE modules.•HarmonicEchoNet shows a 2%–7% accuracy improvement in fetal heart image classification.•HarmonicEchoNet is 10–15 times faster than ConvNeXt, DeiT, and VOLO, with 3.9 ms inference time.•HarmonicEchoNet is highly parameter-efficient, with just 19.9 million parameters.