An Adaptive Synthetic Aperture Method Applied to Ultrasound Tissue Harmonic Imaging

An adaptive beamformer based on dynamic phase coherence factor for pixel-based medical ultrasound imaging

BACKGROUND

Pixel-based beamforming realizes dynamic focusing at the pixel level with a focused beam by assuming that the received signals are composed of spherical pulses. Far-focused pixel-based (FPB) imaging was proposed to avoid artifacts around the focal depth. However, the contrast improvement is limited.

OBJECTIVE

We propose an adaptive weighting method based on dynamic phase coherence factor (DPCF) to improve the image contrast while preserving the speckle pattern.

METHODS

The phase variation is dynamically estimated based on the noise energy proportion of echo signals and it is used to calculate phase coherence weights for suppressing interference and preserving desired signals. A depth-dependent parameter is designed for DPCF to enhance the performance of noise and clutter suppression in the far-field region. We further use the subarray averaging technique to smooth the speckle texture.

RESULTS

The proposed method was evaluated on simulated, phantom experimental, and in vivo data. Results show that, compared with the phase coherence factor (PCF) based method, DPCF respectively leads to average CR improvements by more than 60% and 24% in simulation and experiment, while obtaining an improved speckle signal-to-noise ratio.

CONCLUSIONS

The proposed method is a potentially valuable approach to obtaining high-quality ultrasound images in clinical applications.

Ultrasound far-focused pixel-based imaging using Wiener postfilter scaled by adjustable zero-cross factor

Synthetic aperture (SA) imaging can provide a uniform lateral resolution but an insufficient signal-to-noise ratio (SNR). SA method with bidirectional pixel-based focusing (SA-BiPBF) has the ability to obtain a higher quality image than conventional SA imaging. In this paper, an enhanced SA-BiPBF named full aperture received far-focused pixel-based (FrFPB) is firstly proposed to obtain a high resolution image. An adjustable zero-cross factor scaled Wiener postfilter (AZFsW) is then implemented in FrFPB for improving contrast ratio (CR). The adjustable zero-cross factor is calculated using the polarity of echo signals sequence with an adjustable coefficient σ to estimate the signal coherence, and it is combined with Wiener postfilter to obtain a good capability of noise reduction and background speckle pattern preservation. Simulation and experiments have been conducted to evaluate the imaging performance of the proposed methods. Results show that FrFPB can obviously improve the resolution in comparison with SA-BiPBF, and contrast-to-noise ratio (CNR) and speckle signal-to-noise ratio (sSNR) are retained. In addition, AZFsW can achieve a much higher CR than SA-BiPBF. When σ is 0.6, the CR improvement is 96.7% in simulation, 78.7% in phantom experiment, and 49.2% in in-vivo experiment. To evaluate the imaging performance of AZFsW, coherence factor, conventional Wiener postfilter, and scaled Wiener postfilter are implemented. The imaging results show that when σ is in the range of [0.6, 0.7], AZFsW exhibits a satisfying comprehensive imaging performance.

Spectral Shift Originating from Non-linear Ultrasonic Wave Propagation and Its Effect on Imaging Resolution

An amplitude-dependent downshift in the fundamental wave spectrum of a propagating ultrasonic pulse caused by non-linear wave propagation is described. The effects of non-linearity and the associated downshift on spatial resolution are also studied. The amounts of downshift and spatial resolution are extracted from the numerically simulated beam profile based on the KZK equation. Results for a 25-MHz transducer reveal that non-linear effects can lead to 58% additional downshift in the centre frequency of a pulse compared with a linear case with downshift caused only by attenuation. This additional downshift causes about 50% degradation in axial resolution. However, as the beam becomes narrower from the non-linear effects, the overall effect of non-linearity still leads to improved lateral resolution (≤26%). Therefore, as non-linearity increases with wave pressure, it is concluded that the increase in source pressure improves lateral resolution and degrades axial resolution.

Investigation and Analysis of Ultrasound Imaging Based on Linear CMUT Array

In the next generation of ultrasound imaging systems, Capacitive micromachined ultasonic transducer (CMUT) based on microelectromechanical systems (MEMS) is a promising research direction of transducers, which has wide application prospects. In this paper, based on the study of three imaging methods, including classical phased array (CPA) imaging, classical synthetic aperture (CSA) imaging and phased subarray (PSA) imaging, several different imaging schemes are designed for linear CMUT array, after that the performances of these imaging schemes are compared and analyzed. The effects of the three imaging methods are verified and analyzed based on the linear CMUT array. Through analysis, it is found that the image quality of the classical phased array imaging method is the best, the imaging quality of the above three imaging methods can be effectively improved by adopting the amplitude apodization and dynamic focusing method. The research results in this paper will provide theoretical basis and application reference for the design of ultrasonic imaging system based on linear CMUT array in the future.

Minimum variance based fusion of fundamental and second harmonic ultrasound imaging: Simulation and experimental study

Harmonic imaging is widely used in clinical ultrasound due to its higher resolution in comparison with fundamental mode. However, the low amplitude of harmonic components in this imaging method is a crucial problem, resulting in a high sensitivity to noise, while the fundamental imaging is more robust against noise. To exploit the benefits of both the fundamental and harmonic imaging, we propose a minimum variance (MV)-based adaptive combination of fundamental and harmonic images. The performance of the proposed mixing-together MV (MTMV) beamformer is evaluated on simulated and experimental RF data. The results of the simulated point targets show that in the regions near of point targets, where the desire signals exist, the proposed MTMV beamformer mostly follows the MV-beamformed harmonic image to retain a better resolution. In the regions far from the point targets, where there is just noise, it follows the MV-beamformed fundamental image to benefit from more robustness. Also, the results of the simulated and experimental cyst phantoms indicate that MTMV reduces the background noise level and improves the contrast without compromising the high resolution of the MV-beamformed harmonic image. In the simulated cyst phantom, in comparison to DAS (fundamental), DAS (harmonic), MV (fundamental), MV (harmonic), and wavelet fusion, the image contrast ratio (CR) is increased, in average, about 5.2 dB, 3.5 dB, 1.5 dB, 3.6 dB, and 2.8 dB, respectively. The contrast-to-noise ratio (CNR) is significantly improved; about 59%, 53%, 41%, 37%, and 24%, respectively. In the experimental cyst phantom, these relative improvements are about 6.6 dB, 3.5 dB, 4.2 dB, 2.1 dB, and 3.8 dB for CR, and about 64%, 52%, 25%, 33%, and 33%, for CNR, respectively.

Compressed sensing reconstruction of synthetic transmit aperture dataset for volumetric diverging wave imaging

A high volume rate and high performance ultrasound imaging method based on a matrix array is proposed by using compressed sensing (CS) to reconstruct the complete dataset of synthetic transmit aperture (STA) from three-dimensional (3D) diverging wave transmissions (i.e. 3D CS-STA). Hereto, a series of apodized 3D diverging waves are transmitted from a fixed virtual source, with the ith row of a Hadamard matrix taken as the apodization coefficients in the ith transmit event. Then CS is used to reconstruct the complete dataset, based on the linear relationship between the backscattered echoes and the complete dataset of 3D STA. Finally, standard STA beamforming is applied on the reconstructed complete dataset to obtain the volumetric image. Four layouts of element numbering for apodizations and transmit numbers of 16, 32 and 64 are investigated through computer simulations and phantom experiments. Furthermore, the proposed 3D CS-STA setups are compared with 3D single-line-transmit (SLT) and 3D diverging wave compounding (DWC). The results show that, (i) 3D CS-STA has competitive lateral resolutions to 3D STA, and their contrast ratios (CRs) and contrast-to-noise ratios (CNRs) approach to those of 3D STA as the number of transmit events increases in noise-free condition. (ii) the tested 3D CS-STA setups show good robustness in complete dataset reconstruction in the presence of different levels of noise. (iii) 3D CS-STA outperforms 3D SLT and 3D DWC. More specifically, the 3D CS-STA setup with 64 transmit events and the Random layout achieves ~31% improvement in lateral resolution, ~14% improvement in ratio of the estimated-to-true cystic areas, a higher volume rate, and competitive CR/CNR when compared with 3D DWC. The results demonstrate that 3D CS-STA has great potential of providing high quality volumetric image with a higher volume rate.