Research Of Ultrasound Elastography Algorithms On Multi-range Compression

The pathological changes are often accompanied by changes in tissue stiffness. Many cancers (breast, prostate) are iso-echoic, and hence difficult to detect by ultrasound. For prostate cancer, cirrhosis of the liver, thyroid disease, arterial atherosclerosis and other diseases, an important basis of the diagnosis is the mechanical properties including properties such as stiffness of the tissue.Ultrasound Elastography is an imaging technique that can reflect the tissue mechanical properties. The generated image is called ultrasound elastography, which can reflect distribution of elasticity modulus in the tissue. Ultrasound elastography makes up the deficiencies of the clinical imaging equipment on the mechanical properties imaging, and it provides a new effective diagnosis method for clinicians.Generally speaking, the displacement estimation of tissue directly determines the quality of strain image, but the methods on displacement estimation depend on the amount of tissue compression. For the small amount of compression, the signal decorrelation mainly relies on deformation of the echo signal. However, for the large amount of compression, the signal decorrelation is not only caused by the deformation of echo signal, but also by the lateral displacement of tissue. In order to allow for the influence of signal decorrelation, it is necessary to use different methods to estimate motion of tissue according to different amounts of tissue compression.In this thesis, several motion and strain algorithms based on different amount of tissue compression were studied. For the strain estimate under little compression, a method of1D auto temporal stretching and amplitude correction based on RF signals was proposed. It was proved that this method can compensate the distortion of echo signal and improve the correlation between pre-compression and post-compression signals by experiments of simulation signals and phantom under little compression. For the strain estimate under large compression, a method of2D companding based on analytic signals was introduced. It was proved that this method can compensate the motion of tissue in a large scale and restrain the signal decorrelation through reduce influence of lateral displacement. Finally, the block matching algorithm and optical flow algorithm were applied to estimate displacement of phantom based on the result of2D companding algorithm under large compression. Be verified by FEA, the displacement and strain images can reflect the mechanical properties of the tissue.