Improved Bilateral Filter with Application to Artifact Noise Reduction

Now that stiff tissue deforms less than soft tissue under the same external compression, strain imaging can provide relative stiffness information of biological tissue. However, strain imaging suffers from artifact noise which may come from two dominant sources: de-correlation error and amplitude modulation error. In order to reduce artifacts and improve the quality of ultrasonic strain imaging, this paper proposes a centralized displacement operation based on an improved bilateral filtering (IBF). A new sharpening and smoothing approach was presented in this study, and it contains two important modifications. First, an offset value is introduced to the range filter in the IBF. Second, the processing is in displacement field which the size is smaller than one in strain field. The proposed algorithm has better contrast resolution than the original strain image, and whose performance is better than the conventional bilateral filtering and the spatial compounding method, especially in CNR. Simulation and phantom testing show that the proposed method can improve the quality of ultrasonic strain imaging in terms of tissue SNRe and CNRe values. On the other hand, the presented technique is faster to make for real time than filtering strain field.