Open Access Article
This Open Access Article is licensed under a Creative Commons Attribution 3.0 Unported Licence
Correction: A neural network-based algorithm for high-throughput characterisation of viscoelastic properties of flowing microcapsules
Tao
Lin
a,
Zhen
Wang
b,
Wen
Wang
a and
Yi
Sui
*a
aSchool of Engineering and Materials Science, Queen Mary University of London, London E1 4NS, UK. E-mail: y.sui@qmul.ac.uk
bDepartment of Mechanical Engineering, University College London, London WC1E 6BT, UK
First published on 22nd March 2021
Abstract
Correction for ‘A neural network-based algorithm for high-throughput characterisation of viscoelastic properties of flowing microcapsules’ by Tao Lin et al., Soft Matter, 2021, DOI: 10.1039/d0sm02121k.
The authors wish to include a citation to ‘A high-throughput method to characterize membrane viscosity of flowing microcapsules’ by Tao Lin et al., Physics of Fluids, 2021, DOI: 10.1063/5.0031640.1 They also wish to provide the following clarification for readers:
The work reported in the original article in Soft Matter was carried out in parallel to our recent work in Physics of Fluids (POF). Two different types of neural-network algorithms are reported in the two works that make predictions with different mechanisms. The POF article is based on a convolutional neural network (CNN), while the Soft Matter article combines parts of a CNN with a long short-term memory network (i.e., a hybrid model). A significant advantage/advance of the hybrid model is that it deals with individual instantaneous images of the capsule and builds their temporal connections. It can therefore be much more accurate. However, the CNN (in the POF work) has a merit of being simple.
When analysing the performance of the two methods, similar procedures were followed, and some data of capsule profiles for training and testing the networks were shared. Therefore, some similarities can be seen in the figures of the two papers. In Fig. 3 in the Soft Matter paper and Fig. 5 of the POF paper, and also Fig. 4 of the Soft Matter paper and Fig. 6 of the POF paper, the results regarding performance of the two networks are different by 2–6%. However, these procedures and capsule data are not the main points/results of the papers.
The Royal Society of Chemistry apologises for these errors and any consequent inconvenience to authors and readers.
References
- T. Lin, Z. Wang, R. Lu, W. Wang and Y. Sui, Phys. Fluids, 2021, 33, 011906, DOI:10.1063/5.0031640.
| This journal is © The Royal Society of Chemistry 2021 |