Null subtraction imaging technique for biomedical ultrasound imaging

ultrasound;beamforming;delay and sum;ultrasound imaging;superresolution;apodization

Reeg, Jonathan R ; Oelze ; Michael L.

关键词: ultrasound;    beamforming;    delay and sum;    ultrasound imaging;    superresolution;    apodization;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/90529/REEG-THESIS-2016.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

In ultrasonic imaging, reduction of lateral sidelobes results in an improved image with less distortion and fewer artifacts. In general, apodization is used to lower sidelobes in exchange for increasing the width of the mainlobe and thus decreasing lateral resolution. Null subtraction imaging (NSI) is a nonlinear image processing technique that uses different on-receive apodizations on copies of the same image to reduce sidelobe levels while also dramatically improving lateral resolution. In the NSI technique, three apodization functions are applied on receive.The images created with the different weighting schemes are then added to form a new image with improved sidelobe performance and better lateral resolution compared to conventional apodization. To evaluate the performance of this technique, experimental tasks were performed with an ATS539 phantom containing wire targets to assess lateral resolution and cylindrical anechoic and hyperechoictargets to assess contrast. A 9L4 array was used in the measurements connected to an Ultrasonix RP system with a SonixDAQ.Image reconstruction involved using delay and sum beamforming with apodization. Images were constructed using NSI and compared with a rectangular weighted aperture. In experiments, the lateral resolution was observed to improve by a factor of up to 25 times when compared to rectangular apodization. Image quality was assessed by estimation of lateral resolution (-6-dB receive beamwidth), the mainlobe to sidelobe ratio (MSR) in dB and estimation of the contrast-to-noise ratio (CNR). At f/# = 2 focusing with NSI, the -6-dB beamwidth on receive as measured from a small wire target in the ATS phantom was 0.12lambda. Sidelobes were observed to decrease at each f/# by an average of 29 dB with NSI compared to rectangular apodization. However, the ability to observe the contrast of hyperechoic and anechoic targets was lost when utilizing the NSI scheme, decreasing from -2.13 dB to -0.71 dB for an anechoic target.This could be mitigated by compounding NSI with the B-mode image.