Sua Bae

An effective beamforming algorithm for a GPU-based ultrasound imaging system

In this paper, four beamforming algorithms (i.e., interpolation and phase rotation with pre- and post-filtering, IBF-PRE, IBF-POST, PRBF-PRE and PRBF-POST, respectively) implemented on a high-performance graphics-processing unit (GPU) were presented. Each beamforming method was divided into two kernels consisting of various beamforming and mid-processing blocks and efficiently implemented on a NVIDIAs Computer Unified Device Architecture (CUDA) platform (GeForce GTX560 Ti, NVIDIA, Santa Clara, CA, USA). To evaluate the performance of each method, pre-beamformed radio-frequency (RF) data were captured by a commercial ultrasound machine equipped with a research package (G40, Siemens Healthcare, Mountain View, CA, USA) from a tissue mimicking phantom. The execution time for each beamforming algorithm was measured by using a time stamp produced by a CUDA timer. The IBF-PRE outperforms over other methods (i.e., IBF-POST, PRBF-PRE, PRBF-POST), in terms of execution time, i.e., 7.89 ms vs. 16.19 ms, 21.89 ms, and 10.62 ms, respectively. This result indicates that the IBF-PRE method is suitable for the fully software based ultrasound imaging system.

Color Doppler imaging on a smartphone-based portable US system: Preliminary study

There is increasing interest in the point-of-care US system using smart devices, such as a tablet PC and a smartphone, due to their high accessibility and portability. To meet this growing interest, our group recently has developed a smartphone-based portable US system but it only provided B-mode imaging. In this paper, we present the realization of the Color Doppler (C-mode) imaging on the aforementioned system to provide the functional information such as velocity of blood flow. The smartphone-based portable US system consists of a smart probe and a smartphone. To implement the C-mode processing, we redesign the sequence of the data acquisition on a FPGA (Spartan 6 LX150, Xilinx Inc., USA) of the smart US probe. On the other hand, in the smartphone (Galaxy S5 LTE-A SM-G906, Samsung Electronics Inc., Korea), a real-time C-mode back-end processing is implemented. The performance of the proposed C-mode imaging is evaluated by the phantom study and in vivo study. As a result, the developed system can provide the quantitative information of the velocity and the distribution of the blood flow by B&C-mode images. Also, as the view depth is 4 cm, the proposed C-mode image processing satisfies the requirement of the real-time operation by providing processing time of 22 msec shorter than the data acquisition time (i.e., 34 msec).

Column-based micro-beamformer for improved 2D beamforming using a matrix array transducer

In a 3-D medical ultrasound imaging system, a matrix array probe with 2-D positioning of the elements allows high resolution of ultrasound images due to its capability of two-dimensional dynamic focusing. However, the hundreds (up to thousands) of elements in the matrix array make the fabrication of the transducers and cables challenging. In this paper, to achieve high quality of 3-D ultrasound images with low hardware complexity, we introduce a column-based micro-beamformer (CMB) in which a column in the matrix array is considered as a sub-array and then elevational and lateral beamformings are sequentially conducted in the analog and digital stages, respectively. For the performance evaluation of the proposed beamformer, the beam pattern simulations are conducted and point-spread-functions are obtained. In addition, root-mean-square-errors (RMSEs) in round-trip time delays were measured over the depths. Compared to a conventional micro-beamformer, the CMB produced more tightly focused beam patterns and showed almost equivalent performance to that of fully sampled array. In the near field, the mean RMSEs of the proposed and conventional beamformers were 274 ns and 18.1 ns (improved 93.4% of delay accuracy), respectively.

New shear wave velocity estimation using arrival time differences in orthogonal directions

Shear wave (SW) elasticity imaging is based on the estimation of shear wave speed (SWS) induced by the radiation force of an ultrasound beam. Although the SW initially propagates in the direction perpendicular to the pushing beam, the SW direction changes because of its refraction occurring during traveling in an inhomogeneous medium. The refraction can cause the estimation errors of SWS when the speed is measured along the initial propagation direction. In this paper, we demonstrate that the conventional SWS reconstruction methods yield biased estimates in refractive region. To overcome the problems, also, we propose a novel method in which the differences of SW arrival time in orthogonal directions are used to calculate both the speed and the direction of SW. In the proposed method, SW velocity vectors are estimated and thus their speed can be measured along the propagation direction. Therefore, the estimation errors due to the refraction can be minimized. By the simulations and phantom experiments, we verified that the proposed method can provide more accurate and uniform SWS map than the conventional methods especially in the refractive region.

Real-time realization of adaptive dynamic quadrature demodulation on a gpu-based ultrasound imaging system

In medical ultrasound imaging, the frequency- and depth-dependent attenuation causes the degradation in signal-to-noise ratio (SNR) in quadrature demodulation (QDM). To improve SNR, the adaptive dynamic QDM (ADQDM) method based on a 2nd-order autoregressive (AR) spectral estimation was previously proposed. However, due to its high computational requirements, it is challenging to implement the ADQDM in real time. In this paper, the optimal realization of ADQDM on a GPU-based ultrasound imaging system is presented. To efficiently implement the method, the image is divided into multiple zones, and the center frequency of a receive signal at each zone is independently estimated by using the 2nd-order AR model. The estimated center frequencies are used for dynamic quadrature demodulation. This method was incorporated on the Compute Unified Device Architecture (CUDA) platform and throughputs were measured using a NVIDIAs GTX-560Ti GPU chip. The evaluation was conducted with the beamformed 6144×256 pixel radio-frequency (RF) data which were captured by a commercial ultrasound scanner from the liver of a volunteer. The total execution time for ADQDM is 3.44 ms, which indicates that it can be implemented in real time on a GPU-based medical ultrasound system.

Barker-sequence-modulated golay coded excitation technique for ultrasound imaging

In medical ultrasound imaging, Signal-to-noise ratio (SNR) is an important factor for imaging quality. Coded excitation methods have been widely used such as Golay codes and Barker codes to achieve higher SNR without degradation of axial resolution. The improvement of SNR is determined by code length. However, because conventional Golay codes and Barker codes are limited to specific known sequences, it is only allowed to select one of the known sequences for ultrasound imaging. In this paper, we propose a new Barker-sequence-modulated Golay (BMG) coded excitation technique for ultrasound imaging. The sequences of proposed BMG technique are generated by modulating Golay code with Barker code, which can lead to additional improvement in signal intensity and flexibility for code length compared with conventional Golay code. From the phantom study, the intensity of BMG technique is 28.9dB higher than 1C-pulse. We demonstrate that BMG technique has a higher contrast than the others. To perform a quantitative evaluation, contrast-to-noise ratio (CNR) values were computed from the cyst images. The 6C-BMG method are improved in terms of the CNR value of 4.19.

An optimized plane wave synthetic focusing imaging for high-resolution convex array imaging

High quality ultrasound imaging is required throughout entire imaging depths for abdominal ultrasound. In general, sector scanning with a convex array transducer is performed by using a focused ultrasound beam to obtain a wide field-of-view. However, the conventional focusing (CF) method suffers from the degraded spatial resolution except for the vicinity of the focal depth due to the diffractive propagation of the focused ultrasound. Plane wave synthetic focusing (PWSF) method has been initially proposed for a linear array transducer to give a solution to the diffraction problem of interrogating beam. In this paper, to our best knowledge, we present the first report on the PWSF method for B-mode convex array imaging as well as the optimization strategies for enhancing image quality. The validation study was performed by acquiring data from a tissue-mimicking phantom (040GSE, CIRS Inc., USA) using an ultrasonic research platform (Vantage 128, Verasonics Inc., USA) with a C5-2 convex array transducer. The evaluation results indicate that the PWSF method can be successfully employed to convex array imaging and it exhibits substantially improved image quality compared to that of CF method.

Ultrasound tissue harmonic imaging using nonlinear chirp coded excitation: in vitro study and analysis

The nonlinear chirp tissue harmonic imaging (NC) has been proposed in our group for enhancing the signal-to-noise ratio (SNR) of the harmonic imaging and increasing the frame rate twice compared to the pulse inversion harmonic imaging (PI) by using single transmission for each scanline. Since it has not yet been implemented or evaluated, we demonstrated the feasibility of the NC, evaluated its performance, and compared with other harmonic imaging methods such as the PI without chirp (PIw/oC), the PI with the linear chirp (PI-LC), and the linear chirp (LC). We performed in vitro experiments with an SC1-6 convex array transducer (Alpinion, Republic of Korea) to evaluate and compare the four methods. To our best knowledge, this is the first report on the comparison of LC and NC with a multi-elements channel system. The LC and NC signals were designed with the same bandwidth which instantaneous frequency varies linearly and nonlinearly. The NC was expected to perform better in terms of suppression of the fundamental component than the LC since the overlapped time duration between fundamental and harmonic components in NC is shorter than that of the LC. In in-vitro experiments, NC method suppresses the fundamental component by about 7dB compared with the LC and PI-LC resulting in the better axial resolution.