Ultrasound Image Denoising Using Cascaded Median Filter and Autoencoder

Ultrasound Images are vital to accurately diagnose various ailments in the medical field. Usually., the ultrasound images contain many noises one such is salt and pepper noise. This has a considerable negative effect on the medical diagnosis. In order to remove the salt and pepper noise., this study proposes a cascaded model where we use high efficiency filters such as median filter and then combine a neural network such as DnCNN or Autoencoder to effectively remove noise while retaining important feature information. Experimental outputs show that our denoising method has improved results.