A Joint Convolutional and Spatial Quad-Directional LSTM Network for Phase Unwrapping

TL;DR

This paper presents a novel CNN with Spatial Quad-Directional LSTM for phase unwrapping, improving accuracy under noise and reducing computational time, especially useful with limited training data.

Contribution

It introduces a joint CNN and SQD-LSTM architecture with a custom loss function for effective phase unwrapping, addressing global spatial dependencies.

Findings

Outperforms existing methods under severe noise conditions.

Achieves low normalized RMSE of 1.3% at SNR=0 dB.

Requires less training data and computational time.

Abstract

Phase unwrapping is a classical ill-posed problem which aims to recover the true phase from wrapped phase. In this paper, we introduce a novel Convolutional Neural Network (CNN) that incorporates a Spatial Quad-Directional Long Short Term Memory (SQD-LSTM) for phase unwrapping, by formulating it as a regression problem. Incorporating SQD-LSTM can circumvent the typical CNNs' inherent difficulty of learning global spatial dependencies which are vital when recovering the true phase. Furthermore, we employ a problem specific composite loss function to train this network. The proposed network is found to be performing better than the existing methods under severe noise conditions (Normalized Root Mean Square Error of 1.3 % at SNR = 0 dB) while spending a significantly less computational time (0.054 s). The network also does not require a large scale dataset during training, thus making it ideal for applications with limited data that require fast and accurate phase unwrapping.