MRE data processing pipeline

We present a unified processing pipeline that automates the transformation of raw MRE data into shear modulus containing an optimization-based phase unwrapping method with Dual Data Consistency (Dual-DC) and a TWE-based Neural Network (TWENN) algorithm for modulus inversion. Raw phase images were first unwrapped using an adaptive algorithm employing Dual-DC algorithm. Phase unwrapping and principal component extraction were simultaneously achieved, balancing reconstruction accuracy and noise robustness.

搜集汇总

背景与挑战

背景概述

该数据集提供了一个统一的MRE数据处理管道，能够自动化地将原始MRE数据转换为剪切模量数据，其核心特点包括采用Dual-DC优化相位解包裹方法和TWENN神经网络算法进行模量反演。它通过自适应算法实现相位解包裹和主成分提取，有效平衡了重建精度与噪声鲁棒性。