融合多维超声信号特征的三维TEE成像前端关键理论与技术研究

三维经食管超声心动图（TEE）因其特殊探查位置和优质图像显示，具有明确的临床需求。传统三维TEE成像前端设计基于通用信号处理理论，为了满足成像的实时、宽范围和高质量，成像前端的面积超过换能器标准尺寸、功耗超过人体安全值，成为限制其技术发展的瓶颈。申请人前期通过解决单元级量化的关键技术，已初步实现三维TEE成像前端芯片。在此基础上，本项目将进一步研究体内超声信号特征，构建融合多维超声信号特征的三维TEE成像前端系统设计理论；根据体内超声信号的声衰机理、回波信号的时域稀疏性和频域窄带性，研究动态范围时间增益分解补偿模型、智能识别模型和量化方法、欠带宽缓冲技术；开展融合多维超声信号特征的微型化低功耗三维TEE成像前端芯片设计与验证。本项目的成功实施有望解决三维TEE成像前端面积大功耗高等问题，有助于形成体内超声前端系统设计的共性理论。

3D transesophageal echocardiography (TEE) has clear clinical needs due to its special exploration location and high-quality image. The traditional 3D TEE imaging front-end design is based on general signal processing theory. In order to meet real-time, wide-range and high-quality requirement, the area of the imaging front-end exceeds the standard size of the transducer, and the power consumption exceeds the human safety value, both of which have become technical bottlenecks limiting its development. The applicant has realized a 3D TEE imaging front-end chip by solving the key challenge of element-level quantization. On this basis, this project will further study the characteristics of in vivo ultrasound signals and build a 3D TEE imaging front-end system design theory that incorporates multidimensional features of ultrasound signal. Based on the acoustic attenuation mechanism of in vivo ultrasound signals, the time-domain sparseness and frequency-domain narrowband of echo signals, this project will build dynamic range time gain decomposition compensation model, construct intelligent recognition model and quantization method, and exploit sub-band buffering technology, respectively. Incorporating the multidimensional features, a miniaturized low-power 3D TEE imaging front-end chip will be designed and verified. The successful operation of this project is expected to solve the problems of large front-end area and high power consumption in 3D TEE imaging, and is helpful to form a common theory for in vivo ultrasound front-end system design.