Few-shot Deep Learning for Phase-Amplitude Aberration Correction in Transcranial Focused Ultrasound

Researchers developed a few-shot deep learning model that corrects skull-induced ultrasound distortion in transcranial focused ultrasound using CT scans.

The system predicts per-element phase and amplitude corrections for a 96-element transcranial phased-array transducer.

It uses geometry-aware features, separate phase and amplitude branches, and fine-tuning with only ten target points for new patients.

Across 12 skulls, it closely matched time-reversal simulation performance while running about 2,535 times faster.