Innovative Technique For Ultrasound Image Denoising Based On Iphmf

Abstract

Worldwide Cardiac diseases are considered a major contributor to total deaths. Ultrasound is a frequently utilized system in clinical diagnostics, as it is safe, continuous, non-obtrusive, radiation-free, and more economical than other methods. The aim of this article is to devise an algorithm for denoising of heart ultrasound images.  

Speckle noise is a major issue in ultrasound videos and images. This work proposes a new algorithm, ‘Information Preserving Hybrid Median Filter (IPHMF)’ for speckle noise reduction. The purpose is to improve the preservation of edges, and information capability and make this new filter suitable for ultrasound image denoising in the medical field where information loss cannot be abided. The proposed filter efficiency is tested for the ultrasound videos collected from the hospitals and taken from the open-source repositories.

The proposed method gives results as RMSE: 1.21, PSNR: 46.77, SSIM: 0.97, as compared to LEE (RMSE: 4.10, PSNR: 36.32, SSIM: 0.82), HMF (RMSE: 2.48, PSNR: 40.95, SSIM: 0.92), MHMF (RMSE: 1.61, PSNR: 44.33, SSIM: 0.96). Experimentation shows that IPHMF outperforms, HMF, LEE, and MHMF filters and maintains the balance between speckle suppression and feature preservation based on the performance metrics like RMSE, PSNR, and SSIM.