Motion Estimation For Image Sequence And Ultrasound Elastography Of Artery

The mechanical features of biological tissues have close relation with its physiological and pathological states. Measurements of tissue mechanical characteristics have always been the hotspots of researches in fields of biology, medicine and biomedical engineering. As the combination of mechanics of biological tissues and modem medical imaging techniques, elastography uses biological tissue parameters as its study object and aims to provide useful accessorial information for diagnosis using elasticity map to display the distributive condition of the internal elasticity parameter of the tissue visually. Under some degree, strain distributions reflect the elastic features of tissues. So, although there contained fake images in the final images (including fake images in mechanical aspects and fake images caused by signal processing), it still should be regarded as elastic imaging.Elastography research utilizes medical imaging equipment (ultrosonic equipment, MRI, etc.) to obtain elastic tissue information (one-dimensional or two-dimensional image sequence, this thesis focuses on the latter), have an estimation of tissue displacement and velocity distribution under certain strain action, and reconstruct tissue elasticity modulus distribution based on stress-strain relation and concrete conditions while using full boundary states.Tissue elasticity estimation depends on the measurement of tissue motion information greatly. So, motion estimation for image sequence is the key in elastography. Furthermore, study on movement estimation can also provide foundation for other areas (such as video compression).This study consists of following parts:â‘ This thesis theoretically analyzed the main methods in application as optical flowing method, block matching and elastic matching algorithm. This thesis finally chose elastic matching and created a new motion estimation method after further improvement: elastic matching based on quantum genetic-conjugated gradient hybrid algorithm (QGCGA). It proved that the new method combined the robust overall scanning ability of genetic algorithm and the partial scanning ability of conjugated gradient algorithm, and can process distorted motions as zoom, distortion and circumgyration. It is provided with small match error, high sensitivity for micro-motion, few generations, a great speed of convergence and other merits. Its time-consuming is greatly less than the standard genetic algorithm.â‘¡After summarization for ultrasound elastography fundamentals and atherosclerosis pathology, this thesis developed a perfusion pressurization equipment controlled by computer, carried out ultrasound elastography experiment on artery phantom, and used elastic matching based on QGCGA in estimation for the displacement of outstretched artery phantom, then work out the strain distribution on the transect of the artery wall and phantom wall ,then image it. The result of elastography shows that elastography based on motion estimation for image sequence is provided immense potential in early diagnoses of atherosclerosis.