The homogeneity map method for speckle reduction in diagnostic ultrasound images

This paper proposes a new method to reduce speckle noise in diagnostic ultrasound images by improving the speckle reduction capacity of an edge-sensitive filter. A homogeneity map is generated according to the local statistics of the window formed for each image pixel. The homogeneity map method uses the advantageous properties of a filter with a better smoothing capability than the edge-sensitive filter to increase the image quality of the edge-sensitive filter. We conducted our study using a cyst image simulated by the Field II program, a kidney image, and a real ultrasound image to which we applied two popular filters: SRAD and OBNLM. The results show that our proposed method has better denoising performance than the edge-sensitive filter, without any loss of edges.