AI-Enhanced Ring-Array Ultrasound Computed Tomography

by Wu Qui

Friday Oct 2, 2026, 11:00 AM → 12:00 PM Europe/Berlin

Seminarraum EG + zoom (IPE)

Wu Qui from Huazhong University of Science and Technology in China is visiting IPE. He will give a introduction to ultrasound reconstruction techniques using full wave inversion and artificial intelligence.

Ultrasound Computed Tomography (USCT) is an advanced quantitative imaging modality that reconstructs tissue acoustic parameters such as sound speed and attenuation via full-angle array wavefield acquisition. With radiation-free, standardized and operator-independent imaging capabilities, it shows prominent clinical value in soft-tissue evaluation, especially for breast examination, and has achieved commercial clinical approval. However, traditional USCT suffers from major drawbacks limiting its clinical promotion. Conventional full-waveform inversion relies on heavy iterative computation, leading to low reconstruction speed. Moreover, sampling sparsity and tissue scattering cause severe artifacts, reducing the accuracy of sound speed reconstruction and overall image quality. Artificial intelligence has effectively addressed these challenges in ring-array USCT. This talk focuses on key AI-driven improvements: generative AI models greatly accelerate reflection image reconstruction to support real-time imaging, while physics-informed neural networks substantially enhance the precision of sound speed map reconstruction. Beyond image optimization, AI enables intelligent auxiliary diagnosis, including automatic breast tumor detection for lesion screening and automated muscle volume measurement for standardized soft-tissue assessment. These AI strategies significantly improve USCT practicability, advancing its transition toward accurate, efficient and intelligent clinical imaging.