An efficient fractional-order PDE based image denoising algorithm with optimal adaptive strategy for ultrasound medical image-based diagnostics

Abstract

A fractional partial differential denoising model for ultrasound image and its corresponding finite difference optimization solution algorithm are proposed. The model combines the advantages of the total variational and the fourth-order partial differential equation denoising model, which maintains the edge features and avoids the staircase effect in the smoothing region. In addition, the proposed model employs a dynamic fractional edge detection function with a different order for each pixel point, which is able to adapt to the local texture features of different images. Further, the order optimization objective function is given which incorporates the peak signal-to-noise ratio, structural similarity and mean absolute error. The flower pollination algorithm is proposed to find the optimal order. The proposed model is applied to the real ophthalmic ultrasound image to verify the effectiveness.