| Medical ultrasound imaging is an essential modality in medical diagnosis because it is safe, valid, real-time and low-cost. As the ultrasound image processing algorithms affect the image quality to a large extent and also have a direct impact on the final diagnosis, so the medical ultrasound image processing algorithms have always been a hot research topic in the field of ultrasound imaging.The speckle noise has become a key factor to influence the image quality for the coherence properties of ultrasound, which covers the characteristic of low-contrast tissue, reduces the ability to detect the detailed features for the observers and creates some difficulties for diagnosis. Therefore, it is necessary to filter the speckle noise to improve the ultrasound image quality. Extracting valid information from ultrasound image is a difficult task for even the most experienced physicians, due to the inherent noise present in ultrasound image and low contrast. It needs to convert the images and make them more suitable for human visual system by using a series of technology.In this thesis, the ultrasound image has been divided into unit-region, speckle-including region and edge-including region according to the characteristics of speckle index C,then processes the different regions with different methods. Analysis and tests show that the proposed method can not only improve the performance of denoising but also improve the edge-preserving capability and its performance is superior to traditional Wiener filtering, median filtering, SRAD filtering etc..This thesis proposes an algorithm of nonlinear enhancement. In this algorithm the speckle index is the enhancement factor. Firstly, enhance the edge and improve the contrast by compressing the image reversely, and then improve the brightness by normalizing. Analysis and tests show that the proposed method has enhanced the image. |
