Open AccessRedCap: residual encoder-decoder capsule network for holographic image reconstruction
Author(s) -
Tianjiao Zeng,
Hayden Kwok-Hay So,
Edmund Y. Lam
Publication year - 2020
Publication title -
optics express
Language(s) - Uncategorized
Resource type - Journals
SCImago Journal Rank - 1.394
H-Index - 271
ISSN - 1094-4087
DOI - 10.1364/oe.383350
Subject(s) - computer science , convolutional neural network , residual , artificial intelligence , block (permutation group theory) , holography , deep learning , encoder , artificial neural network , digital holography , computer vision , pattern recognition (psychology) , algorithm , optics , operating system , physics , geometry , mathematics
A capsule network, as an advanced technique in deep learning, is designed to overcome information loss in the pooling operation and internal data representation of a convolutional neural network (CNN). It has shown promising results in several applications, such as digit recognition and image segmentation. In this work, we investigate for the first time the use of capsule network in digital holographic reconstruction. The proposed residual encoder-decoder capsule network, which we call RedCap, uses a novel windowed spatial dynamic routing algorithm and residual capsule block, which extends the idea of a residual block. Compared with the CNN-based neural network, RedCap exhibits much better experimental results in digital holographic reconstruction, while having a dramatic 75% reduction in the number of parameters. It indicates that RedCap is more efficient in the way it processes data and requires a much less memory storage for the learned model, which therefore makes it possible to be applied to some challenging situations with limited computational resources, such as portable devices.