فصلنامه مهندسی پزشکی زیستی
سال سوم شماره 1 (بهار 1388)

In this paper, ultrasonic images are initially deblurred using Gradient method and then the estimations of image and point spread function (PSF) are improved using denoising techniques. For this reason, at first a criterion with appropriate regularizers (that results in preservation of the edges) is defined for the iterative Gradient method, then the estimation of PSF is improved using a denoising technique based on using an anisotropic window around each pixel. The initial estimation of image is also improved using a denoising method in complex wavelet domain that proposes maximum a posteriori (MAP) estimator and local Laplacian prior density function. Using these denoising methods on top of Gradient method causes that our algorithm reduces the visual artifacts and preserves the edges in the deblurred images. Our simulations show that the proposed method in this paper outperforms other methods visually and quantitatively.

Collagen content and its configuration are considered to be among important criteria of healing in tissues. Therefore, developing a method to estimate these factors can benefit physicians in terms of valuable information. In this paper, we examine variation of collagens in tissue mimicking phantoms as well as in vivo tissue taking advantage of applying image processing techniques on ultrasound images of samples. In phantoms, as the base tissue we have used agar-water matrix material and graphite to simulate collagen, respectively. We also have used different concentrations of graphite to simulate different contents of collagen according to attenuation coefficient of ultrasound waves in soft tissue and its correlation with weight ratio of graphite. Experimental and simulation results show that increase in concentration of graphite in phantoms results in higher energy and more contrast level in B-Mode images .... .

Vibroacoustography is a relatively new elasticity imaging method that uses dynamic (oscillatory) radiation force of ultrasound to vibrate the tissue at low frequency (Kilo Hertz). The resulting acoustic emission is recorded with sensitive hydrophone to produce images that are related to the mechanical properties of the tissue. This force is produced by two continuous overlapping ultrasound beams that have a slightly different frequency. Vibroacoustography has been applied to image breast and arteries microcalcification. The lateral resolution of this imaging method is about 0.7mm and its axial resolution is about 12 mm. In this paper two major methods of producing dynamic radiation force, Confocal and X-focal (consists of two concave transducers whose axes cross at their foci at an angle q), are analyzed. A new method for improving axial resolution using short duration pulses is introduced. Simulation results show that we have about 50% improvement in axial resolution using short duration pulses.

In recent years, adaptive beam forming methods have been successfully applied to medical ultrasound imaging, resulting in significant improvement in image quality compared to non-adaptive beam formers. This improvement results from the fact that their weights are chosen based on the priori knowledge of the received data and updated using current statistics of the array signal. Most of the adaptive beam formers presented in the ultrasound imaging literature are based on the minimum variance (MV) beam former, which can improve the imaging resolution while retaining the contrast. It is desirable that the beam former could improve the resolution and contrast, at the same time. To this end, in this paper, we have used temporal averaging besides the conventional spatial averaging to estimate the more accurate covariance matrix. Moreover, we have used the coherence factor weighting combined with MV beam forming to enhance the focusing quality and hence reducing the undesired side lobes. The efficacy of the proposed adaptive beam forming approach is demonstrated via a number of simulated and experimental examples.

The main purpose of this work is introducing a novel method of temperature monitoring using B-Mode Ultrasound digital images. Thermal dependence of sound speed causes a virtual displacement of scatterer particles. The virtual displacement is computed using speckle tracking methods. Horn-Shunck algorithm was applied to a tissue mimicking phantom to measure the virtual displacement. A heating resistor was used in this phantom to generate temperature elevation. The DICOM ultrasound images were acquired using commercial SIMENES ultrasound imaging system with 10MHz linear probe. The accuracy of noninvasive temperature estimation was measured comparing with invasive temperature measurement. The phantom is warmed up to the 8. The mean error of temperature estimation was found to be 0.4°C and peak error 0.9°C. Fast temperature estimation can be achieved using Optical-Flow methods. This Method is a differential based motion estimation method that estimates displacement by calculating the optical pattern changes caused by movements between two frames. Noise sensitivity is the main infirmity of Horn-Schunck method.

Ultrasound is a rapidly growing tool in the field of research, which shows an increasing use in the food industry for both analysis and modification of food products. Quality assessment of agricultural material has an important role in modern agriculture. This study demonstrates the possibility of non-destructive prediction of the main quality indices of the commercial eggs by processing a short ultrasound burst passing through the egg material and calculating the ultrasound phase velocity. For this purpose a set of three hundred samples of commercial eggs (Boris Brown, 33 weeks age) from the first day of egg lying were purchased from a farm and classified in two groups. The first group was kept in the room temperature (22-25°C) while the second group was kept within the refrigerator (4-5°C). 25 eggs were picked every week from each groups (room and refrigerator) were first subjected to the nondestructive ultrasound test at room temperature. Each day, the ultrasound signal is recorded from the eggs first. Then, immediately after that, the air cell, the thick albumen heights, the Haugh unit and the yolk index of the eggs were also determined destructively for comparison purposes. Significant differences at 5% level between the means of the destructive analysis at different days of storage of the eggs were found using ANOVA. Both the Haugh unit and yolk index decreased by time over 5 weeks in storage at room and refrigerator while the air cell height increased. The lower is the Haugh unit for the eggs in the refrigerator the lower is the phase velocity (1573 m/s at first day compared to 1540 m/s after 3 weeks). Similar changes of the phase velocity are found for the eggs in the room temperature (1571 m/s at first day compared to 1514 m/s after 3 weeks).

Increasing number of diagnostic and therapeutic applications of finite amplitude ultrasound in medicine and biology has motivated researchers toward more accurate modeling and more efficient simulation of nonlinear ultrasound regime. One of the most widely used nonlinear models for propagation of 3D diffractive sound beams in dissipative media is the KZK (Khokhlov, Kuznetsov, Zabolotskaya) parabolic nonlinear wave equation. Various numerical algorithms have been developed to solve the KZK equation. Generally, these algorithms fall into one of the three main categories: frequency domain, time domain and combined time-frequency domain. The intrinsic parabolic approximation in the KZK equation imposes limiting accuracy in the solution to the diffraction term of the KZK equation particularly for field points close to the source or in far off-axis region. In this work we developed a novel generalized time domain numerical algorithm to solve the diffraction term of the KZK equation. The algorithm solves the Laplacian operator of the KZK equation in the 3D Cartesian coordinates using novel 5-point Implicit Backward Finite Difference (IBFD) and 5-point Crank-Nicolson Finite Difference (CNFD) techniques. This leads to a more uniform discretization of the Laplacian operator which in turn results in a more accurate solution to the diffraction term in the KZK equation. Comparison between results obtained with the new algorithm and the previously-published data for rectangular ultrasound sources is presented.