Comparative Study Of Despeckle Filtering Of Ultrasound Images

Speckle noise is the main type of noise that affects the quality of ultrasound image.Too much noise will not only obscure the characteristics of the organ in the image, but also reduce the quality and resolution of ultrasound image, thus limiting the further application of the ultrasound image in clinical diagnosis. So, it is necessary to filter the ultrasound image for noise reduction. But researchers agree that speckle noise constains important diagnostic information, especially its distribution and shape which may constains information of the lesions. Based on this consideration, studying the nature of speckle noise and selecting appropriate f iltering algorithms to suppress the image noise while preserving detail features, is very important for accuratee clinical diagnosis, image recognition and segmentation, and feature extraction.Ultrasound image has much speckle noise, low contrast, rich t exture detail, and is sensitive to different types of filtering algorith ms. Most of the recent filtering algorithms are originally applied to other types of i mages, and gradually extended to the ultrasound image processing. Researching the im pact of these algorithms on ultrasound images and evaluating their application in ultrasound image filtering, details preserving and segmentation can provide useful guidance for selecting appropriate filtering algorithms, provide computer-aided diagnosis ideas, and help physicians make right judgmentsThe work of this paper aims to study the effect of different filtering algorithm on ultrasound image, and their application in the field of ultrasound image processing to make scientific evaluation and comparison.To this end, this paper has carried out the following work:(1)This paper has studied the production of speckle noise, its model, its characteristics and the development status of filtering algorithm.(2)This paper has studied 21 kinds of the currently popular filte ring algorithm applied for ultrasound image processing, including algorithm principle, deduction proceduers, performance trait and parameters settings.(3)This paper has studied the application of the Field II ultrasound simulation program and the meaning of each evaluation index.(4)It has done the systematic experimental comparison and analysis for 21 kinds of algorithm applications in the field of speckle noise suppression, image details preserving, pretreatment of image segmention, with the ado ption of liver tumor ultrasound images and synthetic ultrasonic images. What is more, it discusses the application convenience and operation proficiency.In the end, this paper has produced practical conclusion of each algorithm in the three fields with the analysis of experimental chart and data. The result shows that Speckle Reducing Anisotropic Diffusion and Wavelet filtering perform well in three fields. They are appropriate for ultrasound image proc ess to keep details and improve the image quality while at the time reducing noise.