Xiaozhang Zhu

Microwave-Induced Thermal Acoustic Tomography for Breast Tumor Based on Compressive Sensing

Microwave-induced thermal acoustic tomography (MITAT) is an innovative technique to image biomedical tissues based on their electric properties. It has the advantages of both high contrast and high spatial resolution. Image reconstruction method in MITAT is always a critical issue. In this paper, a CS-MITAT (CS: compressive sensing) imaging method is proposed. Compressive sensing (CS) is a recently developed sparse signal representation and analysis framework which handles medical imaging measurements using low sampling rate or increasing imaging quality. The CS-MITAT imaging method applies CS theory to the MITAT for breast tumor imaging. In this method, an over-complete dictionary is established to make sparse measurements in the spatial domain. This treatment greatly saves measurement time. Simulations and experiments with real breast tumor tissues demonstrate the feasibility and effectiveness of the method. Compared with conventional time reversal mirror method which has been used in MITAT research, CS-MITAT provides the same peak signal-to-noise ratio imaging quality by using significantly fewer acoustic sensor positions or scanning times.

Evaluation of Contrast Enhancement by Carbon Nanotubes for Microwave-Induced Thermoacoustic Tomography

Microwave-induced thermoacoustic tomography (MITAT) is a hybrid method which constructs images with ultrasound spatial resolution while exploiting dielectric contrast at microwave frequency. It has great potential in biomedical imaging especially in early breast cancer detection. The detection of early stage breast tumor in MITAT is challenged by the moderate endogenous dielectric contrast between malignant and healthy glandular tissues. In order to overcome this limitation, the performance of using carbon nanotubes (CNTs) as an imaging contrast enhancement agent is evaluated. First, the influences in dielectric and acoustic properties caused by CNTs are measured. Second, based on the measurements and the published data, numerical breast phantom is created and then used to explore the contrast enhancing effect of CNTs for MITAT, by an integrated simulation approach in both electromagnetic and acoustic field. With an experimental MITAT system, the thermoacoustic responses of tissue mimicking materials with different CNTs concentrations are also quantitatively investigated. Finally, the effectiveness of the contrast agent is also validated experimentally by using a MITAT system. The results show that the using of the dielectric contrast agent can effectively enhance the contrast of the MITAT image.

RECONSTRUCTION OF MICROWAVE ABSORPTION PROPERTIES IN HETEROGENEOUS TISSUE FOR MICROWAVE-INDUCED THERMO-ACOUSTIC TOMOGRAPHY

Aiming to e-ciently overcome the acoustic refraction and accurately reconstruct the microwave absorption properties in heterogeneous tissue, an iterative reconstruction method is proposed for microwave-induced thermo-acoustic tomography (MITAT) system. Most current imaging methods in MITAT assume that the heterogeneous sound velocity (SV) distribution obeys a simple Gaussian distribution. In real problem, the biological tissue may have several difierent inclusions with difierent SV distribution. In this case, the acoustic refraction must be taken into account. The proposed iterative method is consisted of an iterative engine with time reversal mirror (TRM), fast marching method (FMM) and simultaneous algebraic reconstruction technique (SART). This method utilizes TRM, FMM and SART to estimate the SV distribution of tissue to solve the phase distortion problem caused by the acoustic refraction efiect and needs little prior knowledge of the tissue. The proposed method has great advantages in both spatial resolution and contrast for imaging tumors in acoustically heterogeneous medium. Some numerical simulation results are given to demonstrate the excellent performance of the proposed method.

Active Adjoint Modeling Method in Microwave Induced Thermoacoustic Tomography for Breast Tumor

To improve the model-based inversion performance of microwave induced thermoacoustic tomography for breast tumor imaging, an active adjoint modeling (AAM) method is proposed. It aims to provide a more realistic breast acoustic model used for tumor inversion as the background by actively measuring and reconstructing the structural heterogeneity of human breast environment. It utilizes the reciprocity of acoustic sensors, and adapts the adjoint tomography method from seismic exploration. With the reconstructed acoustic model of breast environment, the performance of model-based inversion method such as time reversal mirror is improved significantly both in contrast and accuracy. To prove the advantage of AAM, a checkerboard pattern model and anatomical realistic breast models have been used in full wave numerical simulations.

System Development of Microwave Induced Thermo-Acoustic Tomography and Experiments on Breast Tumor

Microwave induced thermo-acoustic tomography (MI- TAT) has become a keen research topic in recent years due to its great potential in early breast cancer detection. A secure and accurate MI- TAT system has been established. Some experiments have been made to demonstrate the performance of the MITAT system. Based on an experiment using phantom, some quantitative features of the system have been obtained. Some imaging experiments with real human breast cancer tissues are performed to demonstrate its efiectiveness and the potential in clinical diagnosis. Images with both high contrast and flne spatial resolution are achieved by using time reversal mirror (TRM) technique in the imaging processing. Moreover, comparisons between the MITAT system result and an ultrasound imaging system result are made. From the comparison, the MITAT system shows its advantages of better contrast over the ultrasound imaging system. The system and the experiments in this paper verify the mechanism of MITAT for breast cancer detection and provide a prototype basis for clinical practice.

Iterative Time-Reversal Mirror Method for Imaging the Buried Object Beneath Rough Ground Surface

An iterative Time-Reversal Mirror (TRM) method is proposed to Detect and Image the buried target beneath ground surface. Unlike the conventional TRM methods which treat the information of the ground as clutters and directly delete them, the iterative TRM imaging method proposed in this paper utilizes the information of rough ground surface as a useful knowledge. The new approach is consisted of two TRM procedures. In the flrst TRM procedure, it aims to image the rough surface where the propagation environment for electromagnetic wave is free space. The second TRM procedure aims to image the buried target. In this step, the information of the rough surface estimated by the flrst TRM procedure will be treated as newly updated propagation environment. Then conventional TRM is applied to image the buried target. By applying this iterative TRM method, the information of the rough ground can be well considered in the whole TRM procedure. Numerical simulations prove that this method performs signiflcantly better image contrast comparing with the results obtained by using conventional TRM. 4{5dB improvement on the imaging SNR has been achieved. Furthermore, the target can be located more accurately.

Block based compressive sensing method of microwave induced thermoacoustic tomography for breast tumor detection

Microwave induced thermoacoustic tomography (MITAT) is a developing non-ionized technique which has great potential in early breast tumor detection. In our previous work, an imaging method, CS-MITAT, was proposed, which applied the compressive sensing theory in MITAT and achieved a good image. The method converts a signal model into an unconstrained optimization problem with l1 norm regularization, which only exploits the spatial sparsity of targets. In this paper, based on the block sparsity of thermoacoustic signals and target distribution in MITAT, the signals to be detected can be grouped into several blocks and the summation of l2 norm regularization is used to replace the l1 norm regularization of the CS-MITAT method. The combination of l2 and l1 norm regularizations helps the aggregation of nonzero elements which are accumulated in blocks. A priori structural constraint is added to form a more realistic signal model which can improve the image quality. Compared with the conventional approach of tim...

A prototype system of microwave induced thermo-acoustic tomography for breast tumor

Microwave-induced thermo-acoustic tomography (MITAT) is an innovative technique for tumors detection. Due to there has high contrast in terms with permittivity and electrical conductivity of tumor versus normal tissue, even if the tumor still in the early phase it can be imaged clearly. For the proposed MITAT system, low energy microwave pulses are used as the irradiating signals, while the received signals are ultrasound, high contrast and high resolution images can be obtained. After some theoretical research and basic fundamental experiments, the first prototype of experimental system is designed and built. It includes the microwave radiator, the arrayed sensor bowl, the circular scanning platform, the system controller and the signal processor. Based on the experimental results using this integral MITAT clinic system, the images contrast can be reached higher than 383:1; while the sub-millimeter special resolution is obtained for a 1cm3 scale tumor mimic.

Hierarchical dictionary compressive sensing (HDCS) method in microwave induced thermal acoustic tomography

Abstract Aiming to reduce the reconstruction time and enhance the image quality of microwave induced thermal acoustic tomography (MITAT), a new image reconstruction method named HDCS-MITAT (HDCS: hierarchical dictionary compressive sensing) is proposed. Different from the recently demonstrated CS-MITAT (CS: compressive sensing) imaging method in which only one level dictionary is applied, hierarchical dictionaries are used in the HDCS-MITAT. In this method, the dictionaries with different spatial resolutions are constructed which constitute a hierarchical structure. During the image reconstructions, first the coarsest level dictionary is utilized to roughly estimate the position of the targets in the original image domain. A reduced interested image domain can be set based on this estimation. Then the next level dictionary which has higher resolution than the above level is applied to further estimating the position of the targets and so on. Finally, the finest level dictionary is used to reconstruct the image of the targets. Compared with the CS-MITAT, this HDCS-MITAT has much less computational time and better image quality. The effectiveness of the method has been validated through some simulations and real breast tumor experiments.

Resolving ambiguities in DOA estimation by optimizing the element orientations

A method of resolving ambiguities in direction-of-arrival (DOA) estimation for sparse linear arrays with directional antennas is reported. The DOA estimation ambiguity comes from the linear dependence among the array manifold vectors of different DOAs. Optimizing the element orientations can break down the linearity. The element orientations are optimized by using particle swarm optimization (PSO) algorithm with the Weiss-Weinstein bound (WWB). This method is demonstrated through simulations of a 5-element collinear array of dipole antennas.