請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/22573
標題: | 適用於高速超音波成像系統的低複雜度可適性波束成像器引擎 Low-Complexity Adaptive Beamformer Engine in High Frame Rate Ultrasound Imaging System |
作者: | Ming-Chia Tsai 蔡銘嘉 |
指導教授: | 吳安宇 |
關鍵字: | 超音波,波束成像器,數位訊號處理,多核心系統, Ultrasound,Beamformer,DSP,Multi-core system, |
出版年 : | 2010 |
學位: | 碩士 |
摘要: | Medical imaging system is to provide a visual representation of anatomical structures, blood flow velocity, and other diagnostic information. Currently, the mainly three applications in ultrasound system are B-mode image processing, color Doppler, or spectral Doppler. B mode image represents the power of the received echo. The Doppler mode is used for blood flow estimation and cardiac muscle observation. In this thesis, we focus on B mode imaging. Since the natural characteristics of the propagation wave, beamforming thus play a vital role in array signal processing, whose functionality is to focus the received echoes. Beamforming can be performed by real aperture and synthetic aperture with corresponding ways of delay calculation to focus echoes. The oldest but the most popular method is to delay and sum for received echoes alignment, which we adopt in this work. Real aperture use full array, which consumes more power and computation complexity to form an image. Synthetic aperture uses less hardware overhead in trade of image quality. High frame rate imaging system (HFR) is just composed of synthetic transmit aperture and broad transmit beams (plane wave, defocused beams, and diffraction beams).
The techniques of digital signal processing can overcome this problem by weighting adjustment on channels. Methods of constant weighting are often used to beam shaping; however, adaptive methods are also applied to medical signal processing recently. For example, CFMVDR aims to enhance terrible image quality in HFR system. The problem is the high complexity of weighing calculation. In this thesis, we propose a new method of low complexity adaptive imaging. This method has little hardware overhead without performance loss. We have built the proposed model on Matlab to test the efficacy by both synthetic data using Field II and clinical data of breast from female patients. Still, we implement the proposed adaptive beamforming on CUDA (Compute Unified Device Architecture, new platform of multi-core computing) and propose some methods of modification. In the assistance of multi-core platform, the proposed adaptive imaging can be proven to approach the requirement of real time application, which is ever published. Through this thesis, we offer the new concepts combined with adaptive algorithms and implementations for future research in this domain. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/22573 |
全文授權: | 未授權 |
顯示於系所單位: | 電子工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-99-1.pdf 目前未授權公開取用 | 5.72 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。