Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/91716
Title: | 行列式陣列波束成型對於流速計算準確度之影響 Effects of Beamforming on Flow Estimation with Row Column Addressed Arrays |
Authors: | 黃資芸 Tzu-Yun Huang |
Advisor: | 李百祺 Pai-Chi Li |
Keyword: | 超音波三維血流影像,波束成型,旁瓣抑制,平面波複合,K空間濾波, 3D ultrasound flow imaging,beamforming,sidelobe suppression,plane wave compounding,K-space filtering, |
Publication Year : | 2024 |
Degree: | 碩士 |
Abstract: | 超音波血流影像技術在近年來由二維發展至三維,能夠提供更多的空間訊息、呈現血管的立體結構以及更全面的血流動力學資訊。血流影像的成像品質提高對於臨床診斷至關重要,有助於提高診斷的精確性,並能夠早期發現潛在的血管問題。然而,在使用行列式陣列的三維超音波影像中,雖然硬體以及計算複雜度得到降低,但減少的傳感器元件和較長孔徑將會導致嚴重的旁瓣雜訊,進而影響成像品質。本論文認為高強度的旁瓣雜訊可能影響血流流速的準確度。為了探討這一影響,研究使用不同的變跡函數調整旁瓣訊號的強弱,並使用彩色都卜勒計算血流流速,觀察旁瓣訊號強弱與流速量測準確度之間的關係。結果顯示,旁瓣訊號越弱的模擬影像具有較高的整體流速準確度,其平均流速越接近仿體模擬設置的0.3m/s,而旁瓣訊號較強的模擬影像則會導致流速較低且分布不均勻的現象,平均流速的誤差將增加約0.14m/s,由結果可知旁瓣訊號的強弱會影響血流流速偵測的準確度。論文進一步探討了三種針對行列式陣列設計的波束成型方法,包括平面波複合搭配延遲加總波束合成方法、行列式幀相乘與加總波束成型方法,以及K空間濾波重建全採樣陣列資料方法。透過旁瓣抑制指標和彩色都卜勒影像的分析,比較這三種方法對於旁瓣訊號抑制效果以及血流測量準確度的影響。結果顯示,這三種成像品質提升方法均能成功抑制旁瓣,並提高血流流速量測之準確度,平均流速的誤差從0.13m/s減少到0.08m/s。除此之外,比較使用有限脈衝響應高通濾波器以及奇異值分解濾波方法對於過濾組織訊號之能力,結果顯示奇異值分解濾波方法減少了約0.2m/s的誤差。未來在血流影像的波束成型序列設計中,旁瓣抑制應被視為一項重要的考量因素。 Ultrasound flow imaging technology has evolved from 2D to 3D in recent years, offering increased spatial information, depicting vascular three-dimensional structures, and providing comprehensive hemodynamic insights. The improvement in imaging quality of blood flow is crucial for clinical diagnosis, enhancing diagnostic accuracy, and facilitating the early detection of potential vascular issues. However, in the context of 3D ultrasound imaging using row-column addressed (RCA) 2D arrays, while hardware and computational complexity are reduced, the decrease in transducer elements and longer aperture size result in significant sidelobe artifacts that impact imaging quality. This paper posits that high-intensity sidelobe noise may affect the accuracy of blood flow velocity measurements. This study explores this impact by utilizing different apodization functions to adjust the strength of sidelobe signals and use color Doppler to calculate blood flow velocity, observing the relationship between sidelobe signal strength and flow measurement accuracy. Results indicate that images with weaker sidelobe signals exhibit higher velocity accuracy, with an average velocity closer to the simulated setting of 0.3 m/s. In contrast, images with stronger sidelobe signals result in lower and uneven flow velocity, with an average velocity error of approximately 0.14 m/s. These findings highlight the impact of sidelobe signal strength on flow velocity detection accuracy. The paper further explores three beamforming methods tailored for RCA array: plane wave compounding with delay and sum beamforming, row-column frame multiply and sum beamforming, and imaging level enhancement and k-space filtering method. Through analysis using sidelobe suppression indices and color Doppler imaging, the study compares the effectiveness of these three methods in sidelobe signal suppression and blood flow measurement accuracy. Results demonstrate that all three image quality enhancement methods successfully suppress sidelobes and improve the accuracy of flow velocity measurements. The average velocity error is reduced from 0.13 m/s to 0.08 m/s. The study also compared the capabilities of using FIR high-pass filter and singular value decomposition filtering methods in filtering tissue signals. The results show that the singular value decomposition filtering method reduced the average velocity error by 0.2 m/s. In future designs of beamforming sequences for blood flow imaging, sidelobe suppression should be considered a crucial factor. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/91716 |
DOI: | 10.6342/NTU202400484 |
Fulltext Rights: | 未授權 |
Appears in Collections: | 生醫電子與資訊學研究所 |
Files in This Item:
File | Size | Format | |
---|---|---|---|
ntu-112-1.pdf Restricted Access | 10.33 MB | Adobe PDF |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.