The Measurement And Characterization Of The Elastical Properties Of Soft Tissues

The technology of ultrasonic tissue characterization is a hotspot in the field ofbiomedical engineering after 70's of last century, which mainly extracts soft-tissueelasticity parameters by processing the signals of ultrasound echo from soft tissueunder certain loads, and serves as a method for obtaining acousto-elasticity pictures.The key of elasticity estimation is estimating the strains under different loads.Currently, many algorithms with quite different effects have been proposed. In spiteof the significant progresses in theoretical researches, no report on portable clinicinstruments has been announced in our country. Therefore, it is necessary to develop aportable tester for measuring soft-tissue elasticity parameters based on ultrasound andto prove the feasibility of the technology. The design of the portable tester, including the transmission of ultrasound, thehardware and software debugging of the DAQ circuit has been accomplished in thethesis. In order to prove the effectiveness of various estimation algorithms, sometypical algorithms are compared. Then, soft-tissue elasticity parameters are obtainedby using the proposed algorithms and the elasticity parameters revealed during thepathological changes of the soft-tissues are compared. The DSP is used as the CPU in the design and development of portable tester,and they are developed by CPLD technology. By using DMA, the real-time DAQ ofultrasound and load signals is obtained. An ultrasonic diagnostic system formeasuring elastical properties of soft tissues based on PC has been developed byutilizing the existing resources in our lab. The system can deal with the data sampling,conditioning, synchronous acquisition and display of the ultrasonic and force signals.The cross-correlation algorithm, which estimates the strain by measuring the timedifference between ultrasound echoes before and after compression for obtainingmaximum cross-correlation coefficient, is used. Some improvements in maximumsearching have been made in the thesis. Besides, a novel algorithm has been proposedto reduce the error from cross-correlation. Finally, an analysis and comparison ofexperiment data are presented.