Research On Algorithm Simulation Of Biological Tissues Strain Estimation Based On Ultrasound Elastography

The elastic properties of soft tissues depend on their molecular building blocks,and on the microscopic and macroscopic structural organization of these blocks.The changes in tissues’ elastic properties are correlated with pathological changesas well. So there are some differences in stiffness between the lesion and thesurrounding normal tissue. Be imposed on with an incentive, biological soft tissuehas a deformation, so we can obtain the pre-and postcompression RF signals andestimate the displacement distribution of tissue, and then get the strain distributionand elastic modulus distribution of biological soft tissue by means of signalprocessing and cross-correlation. Based on the reflection-type ultrasoundelastography system, the thesis put forward a1-D axial strain estimation method,which can accurately estimate the elastic modulus (or hardness) distribution ofbiological soft tissue. The main work of the thesis is as follows:Firstly, it introduces one fundamental mechanics law of biological soft tissuein ultrasound elastography–Hooke’s law, and studies mechanical properties ofbiological soft tissue and that elastography is applied in the tissue so that simplify itto build the mechanical model of biological tissue in elastography. It introduces theprinciple of ultrasonic elastography and some most common elastographies andultrasound elastography system and its classifications based on the principle ofultrasonic elastography of quasi-static or static compression, and studies on theimaging principle of reflection-type ultrasonic elastography system compared withtransmission-type ultrasonic elastography system.Secondly, it studies on cross-correlation algorithm, a classical algorithm forultrasonic elastography, and traditional time-domain cross-correlation algorithmbased on cross-correlation algorithm. Considering the deficiencies of traditionalmethods, the thesis puts forward a method of computing scaling factor forestimating strain of tissue. It designs a method calculating the numbers ofinterpolation points. The method is able to get a correct allocation plan of scalingfactor in accord with the elastic modulus of tissue, but its precision is too small,only1%. Therefore, on that basis it optimizes to calculating the scaling factor toestimate the tissue strain.Thirdly, the thesis summarizes and proposes preferences method and selectinginterpolation method to analyze elastography algorithms. Studying that differentwindow lengths and overlaps influence on elastography algorithms, do manysimulation experiments and then fix window length as100and overlap as0.95;compare and analyze linear interpolation, cubic interpolation, spline interpolation and so on, select linear interpolation to stretch signals, for it has short performanceperiod and high precision. In order to improve the speed of simulation test,according to the structure characteristics of the imaging algorithm, I optimize themethod to the biological image processing platform developed by ourselveslaboratory and cut short performance period greatly.Finally, I simulated the methods mentioned in the thesis and diddemonstrations and discussions on the simulation results. Then I summarized thesubject and did some prospects to do next researches.