File Download
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1364/OE.383350
- Scopus: eid_2-s2.0-85079342100
- PMID: 32121718
- WOS: WOS:000514575500045
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: RedCap: residual encoder-decoder capsule network for holographic image reconstruction
Title | RedCap: residual encoder-decoder capsule network for holographic image reconstruction |
---|---|
Authors | |
Keywords | Computational resources Data representations Digit recognition Digital holographic reconstruction Encoder-decoder |
Issue Date | 2020 |
Publisher | Optical Society of America: Open Access Journals. The Journal's web site is located at http://www.opticsexpress.org |
Citation | Optics Express, 2020, v. 28 n. 4, p. 4876-4887 How to Cite? |
Abstract | 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. |
Persistent Identifier | http://hdl.handle.net/10722/287883 |
ISSN | 2023 Impact Factor: 3.2 2023 SCImago Journal Rankings: 0.998 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | ZENG, T | - |
dc.contributor.author | So, HKH | - |
dc.contributor.author | Lam, EYM | - |
dc.date.accessioned | 2020-10-05T12:04:38Z | - |
dc.date.available | 2020-10-05T12:04:38Z | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | Optics Express, 2020, v. 28 n. 4, p. 4876-4887 | - |
dc.identifier.issn | 1094-4087 | - |
dc.identifier.uri | http://hdl.handle.net/10722/287883 | - |
dc.description.abstract | 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. | - |
dc.language | eng | - |
dc.publisher | Optical Society of America: Open Access Journals. The Journal's web site is located at http://www.opticsexpress.org | - |
dc.relation.ispartof | Optics Express | - |
dc.rights | Optics Express. Copyright © Optical Society of America: Open Access Journals. | - |
dc.rights | © 2020 [Optical Society of America]. Users may use, reuse, and build upon the article, or use the article for text or data mining, so long as such uses are for non-commercial purposes and appropriate attribution is maintained. All other rights are reserved. | - |
dc.rights | This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. | - |
dc.subject | Computational resources | - |
dc.subject | Data representations | - |
dc.subject | Digit recognition | - |
dc.subject | Digital holographic reconstruction | - |
dc.subject | Encoder-decoder | - |
dc.title | RedCap: residual encoder-decoder capsule network for holographic image reconstruction | - |
dc.type | Article | - |
dc.identifier.email | So, HKH: hso@eee.hku.hk | - |
dc.identifier.email | Lam, EYM: elam@eee.hku.hk | - |
dc.identifier.authority | So, HKH=rp00169 | - |
dc.identifier.authority | Lam, EYM=rp00131 | - |
dc.description.nature | published_or_final_version | - |
dc.identifier.doi | 10.1364/OE.383350 | - |
dc.identifier.pmid | 32121718 | - |
dc.identifier.scopus | eid_2-s2.0-85079342100 | - |
dc.identifier.hkuros | 314913 | - |
dc.identifier.volume | 28 | - |
dc.identifier.issue | 4 | - |
dc.identifier.spage | 4876 | - |
dc.identifier.epage | 4887 | - |
dc.identifier.isi | WOS:000514575500045 | - |
dc.publisher.place | United States | - |
dc.identifier.issnl | 1094-4087 | - |