Study On The Methods For Speckle Noise Reduction In Ultrasound Images

Ultrasonic imaging has been widely applied and playes a powerfulrole in medical diagnosing, industrial nondestructive detecting, marineprospecting and underwater warfare. Ultrasound images are perturbed byspeckle noise,which is difficult for manual reading and computeridentifying.As a tough issue, it has restricted the development ofUltrasonic imaging. Studying on the methods for speckle noise reductionin ultrasound images plays an important role in ultrasound imagesapplying. For this purpose，the paper studys on the speckle noisereduction.At first, the present state of research on speckle reduction isintroduced after consulting a huge mass of reference and the studydirection is fixed.Study on median filter, Wiener filter, Lee filter andanisotropic diffusion based filter, compare the speckle reduction results in real-time, application condition and automation degree.Then, make a description of theoretical knowledge about wavelettransform and application in image process.The paper presents animproved BayesShrink method, which is better than soft thresholdmethod in preserving wavelet coefficients owing to its continuity. Thenew method was employed in liver ultrasound image and sonarimage.The results of the experiments show that the method proposed isbetter than median filtering, Wiener filtering and soft BayesShrinkthreshold method in vision and embrace objective denosing datas.At last, study on BayesShrink speckle reduction andDT-CWT-Wiener filtering based on the theory of DT-CWT and itsexcellent features.The method of BayesShrink on DT-CWT is excellentin speckle reduction except for preserving edges and details.TheDT-CWT-Wiener filtering uses the Wiener filter in every scale and the adjacent wavelet coeffiecients are included.As a result, the results of theexperiments show that it is better in speckle noise reduction and inpreserving edges and details.