H-ProMed: Ultrasound image segmentation based on the evolutionary neural network and an improved principal curve

作者：

Highlights：

• A hybrid method (H-ProMed) is proposed for accurate and robust prostate segmentation in TRUS images.

• An optimized closed polygonal segment method is newly proposed to obtain data sequences.

• An improved dynamic storage-based differential evolution method is newly used to assist in finding the optimal caputo fractional-order backpropagation training network.

• A smooth mathematical model of the prostate contour is developed to express the prostate contour.

摘要

•A hybrid method (H-ProMed) is proposed for accurate and robust prostate segmentation in TRUS images.•An optimized closed polygonal segment method is newly proposed to obtain data sequences.•An improved dynamic storage-based differential evolution method is newly used to assist in finding the optimal caputo fractional-order backpropagation training network.•A smooth mathematical model of the prostate contour is developed to express the prostate contour.

论文关键词：Accurate prostate segmentation,Transrectal ultrasound,Principal curve,Optimized closed polygonal segment method,Evolutionary neural network,Interpretable mathematical model

论文评审过程：Received 12 May 2021, Revised 22 June 2022, Accepted 4 July 2022, Available online 8 July 2022, Version of Record 12 July 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108890