AI breakthrough improves ultrasound image quality by 30% using wavefield correlation

Researchers have developed a wavefield correlation (WFC) approach that significantly improves ultrasound image quality by optimizing sound speed estimation. The method uses auto-differentiation software and gradient descent optimization to reduce estimation errors in media with clutter and large sound speed changes. Compared to traditional delay-and-sum methods, WFC decreases sound speed estimation error and improves image resolution and contrast by up to 30% in simulated, phantom, and in vivo tests.