A novel despeckling approach for ultrasound images using adaptive OBNLM filter

The Optimize Bayesian Non-Local Mean Filter (OBNLM) provides a very strong tool for despeckling in Ultrasound. However, some parameters of this filter depend on the input (noise) and they are difficult to adjust. This article generates a denoising solution using the adaptive OBNLM filter in combination with the Binary Bat Algorithm (BBA) and on the no-reference Q-Metric (BBA-OBNLM). The proposed filter can despeckle noise without the need for reference images and still keep the image details, edges and textures in good condition. Furthermore, in this article, we have also carried out some simulations with images which are added speckle noise with different variances to demonstrate the performance of the proposed method superior to previous publications. © 2019, Springer Nature Switzerland AG.