Skip to Main Content
Adaptive Anisotropic Filtering using local structure tensor analysis has been used for multi-dimensional ultrasound image enhancement. This paper presents a modified version of the existing framework for adaptive filtering mechanism with the view of application to ultrasound images. The goal of this work is to enhance and preserve important, typically anisotropic, image structures while suppressing high-frequency noise. This filtering technique facilitates user interaction and direct control over high frequency contents of the signal. Local structure analysis is performed based on tensor estimation with an optimized set of spherical harmonic filters. The optimized filters significantly reduce the overall computational cost with minimum risk on accuracy. In this paper the algorithm is applied to phantom and in vivo ultrasound data. The performance measure of the algorithm is evaluated in terms of Contrast to Noise Ratio (CNR). Experimental results show that a good level of speckle reduction along with structure enhancement can be achieved in the adaptive filtered ultrasound images.