Intelligent Speckle Reducing Anisotropic Diffusion Algorithm for Automated 3-D Ultrasound Images

Jun Wu,Yuanyuan Wang,Jinhua Yu,Xinling Shi,Junhua Zhang,Yue Chen,Yun Pang
DOI: https://doi.org/10.1364/josaa.32.000248
2015-01-01
Abstract:A novel 3-D filtering method is presented for speckle reduction and detail preservation in automated 3-D ultrasound images. First, texture features of an image are analyzed by using the improved quadtree (QT) decomposition. Then, the optimal homogeneous and the obvious heterogeneous regions are selected from QT decomposition results. Finally, diffusion parameters and diffusion process are automatically decided based on the properties of these two selected regions. The computing time needed for 2-D speckle reduction is very short. However, the computing time required for 3-D speckle reduction is often hundreds of times longer than 2-D speckle reduction. This may limit its potential application in practice. Because this new filter can adaptively adjust the time step of iteration, the computation time is reduced effectively. Both synthetic and real 3-D ultrasound images are used to evaluate the proposed filter. It is shown that this filter is superior to other methods in both practicality and efficiency.
What problem does this paper attempt to address?