Standard

Intelligent Identification of MoS(2)Nanostructures with Hyperspectral Imaging by 3D-CNN. / Li, Kai-Chun; Lu, Ming-Yen; Hong Thai Nguyen et al.

In: Nanomaterials, Vol. 10, No. 6, 1161, 06.2020.

Research output: Contribution to journalArticlepeer-review

Harvard

Li, K-C, Lu, M-Y, Hong Thai Nguyen, Feng, S-W, Artemkina, SB, Fedorov, VE & Wang, H-C 2020, 'Intelligent Identification of MoS(2)Nanostructures with Hyperspectral Imaging by 3D-CNN', Nanomaterials, vol. 10, no. 6, 1161. https://doi.org/10.3390/nano10061161

APA

Li, K-C., Lu, M-Y., Hong Thai Nguyen, Feng, S-W., Artemkina, S. B., Fedorov, V. E., & Wang, H-C. (2020). Intelligent Identification of MoS(2)Nanostructures with Hyperspectral Imaging by 3D-CNN. Nanomaterials, 10(6), [1161]. https://doi.org/10.3390/nano10061161

Vancouver

Li K-C, Lu M-Y, Hong Thai Nguyen, Feng S-W, Artemkina SB, Fedorov VE et al. Intelligent Identification of MoS(2)Nanostructures with Hyperspectral Imaging by 3D-CNN. Nanomaterials. 2020 Jun;10(6):1161. doi: 10.3390/nano10061161

Author

Li, Kai-Chun ; Lu, Ming-Yen ; Hong Thai Nguyen et al. / Intelligent Identification of MoS(2)Nanostructures with Hyperspectral Imaging by 3D-CNN. In: Nanomaterials. 2020 ; Vol. 10, No. 6.

BibTeX

@article{6ab737810cee44418510943318c5202c,
title = "Intelligent Identification of MoS(2)Nanostructures with Hyperspectral Imaging by 3D-CNN",
abstract = "Increasing attention has been paid to two-dimensional (2D) materials because of their superior performance and wafer-level synthesis methods. However, the large-area characterization, precision, intelligent automation, and high-efficiency detection of nanostructures for 2D materials have not yet reached an industrial level. Therefore, we use big data analysis and deep learning methods to develop a set of visible-light hyperspectral imaging technologies successfully for the automatic identification of few-layers MoS2. For the classification algorithm, we propose deep neural network, one-dimensional (1D) convolutional neural network, and three-dimensional (3D) convolutional neural network (3D-CNN) models to explore the correlation between the accuracy of model recognition and the optical characteristics of few-layers MoS2. The experimental results show that the 3D-CNN has better generalization capability than other classification models, and this model is applicable to the feature input of the spatial and spectral domains. Such a difference consists in previous versions of the present study without specific substrate, and images of different dynamic ranges on a section of the sample may be administered via the automatic shutter aperture. Therefore, adjusting the imaging quality under the same color contrast conditions is unnecessary, and the process of the conventional image is not used to achieve the maximum field of view recognition range of similar to 1.92 mm(2). The image resolution can reach similar to 100 nm and the detection time is 3 min per one image.",
keywords = "hyperspectral imagery, deep learning, 3D-CNN, MoS2, automated optical inspection, MONOLAYER MOS2, LAYER MOS2, SINGLE-LAYER, MIXED PIXEL, EVOLUTION, GROWTH",
author = "Kai-Chun Li and Ming-Yen Lu and {Hong Thai Nguyen} and Shih-Wei Feng and Artemkina, {Sofya B.} and Fedorov, {Vladimir E.} and Hsiang-Chen Wang",
year = "2020",
month = jun,
doi = "10.3390/nano10061161",
language = "English",
volume = "10",
journal = "Nanomaterials",
issn = "2079-4991",
publisher = "MDPI AG",
number = "6",

}

RIS

TY - JOUR

T1 - Intelligent Identification of MoS(2)Nanostructures with Hyperspectral Imaging by 3D-CNN

AU - Li, Kai-Chun

AU - Lu, Ming-Yen

AU - Hong Thai Nguyen, null

AU - Feng, Shih-Wei

AU - Artemkina, Sofya B.

AU - Fedorov, Vladimir E.

AU - Wang, Hsiang-Chen

PY - 2020/6

Y1 - 2020/6

N2 - Increasing attention has been paid to two-dimensional (2D) materials because of their superior performance and wafer-level synthesis methods. However, the large-area characterization, precision, intelligent automation, and high-efficiency detection of nanostructures for 2D materials have not yet reached an industrial level. Therefore, we use big data analysis and deep learning methods to develop a set of visible-light hyperspectral imaging technologies successfully for the automatic identification of few-layers MoS2. For the classification algorithm, we propose deep neural network, one-dimensional (1D) convolutional neural network, and three-dimensional (3D) convolutional neural network (3D-CNN) models to explore the correlation between the accuracy of model recognition and the optical characteristics of few-layers MoS2. The experimental results show that the 3D-CNN has better generalization capability than other classification models, and this model is applicable to the feature input of the spatial and spectral domains. Such a difference consists in previous versions of the present study without specific substrate, and images of different dynamic ranges on a section of the sample may be administered via the automatic shutter aperture. Therefore, adjusting the imaging quality under the same color contrast conditions is unnecessary, and the process of the conventional image is not used to achieve the maximum field of view recognition range of similar to 1.92 mm(2). The image resolution can reach similar to 100 nm and the detection time is 3 min per one image.

AB - Increasing attention has been paid to two-dimensional (2D) materials because of their superior performance and wafer-level synthesis methods. However, the large-area characterization, precision, intelligent automation, and high-efficiency detection of nanostructures for 2D materials have not yet reached an industrial level. Therefore, we use big data analysis and deep learning methods to develop a set of visible-light hyperspectral imaging technologies successfully for the automatic identification of few-layers MoS2. For the classification algorithm, we propose deep neural network, one-dimensional (1D) convolutional neural network, and three-dimensional (3D) convolutional neural network (3D-CNN) models to explore the correlation between the accuracy of model recognition and the optical characteristics of few-layers MoS2. The experimental results show that the 3D-CNN has better generalization capability than other classification models, and this model is applicable to the feature input of the spatial and spectral domains. Such a difference consists in previous versions of the present study without specific substrate, and images of different dynamic ranges on a section of the sample may be administered via the automatic shutter aperture. Therefore, adjusting the imaging quality under the same color contrast conditions is unnecessary, and the process of the conventional image is not used to achieve the maximum field of view recognition range of similar to 1.92 mm(2). The image resolution can reach similar to 100 nm and the detection time is 3 min per one image.

KW - hyperspectral imagery

KW - deep learning

KW - 3D-CNN

KW - MoS2

KW - automated optical inspection

KW - MONOLAYER MOS2

KW - LAYER MOS2

KW - SINGLE-LAYER

KW - MIXED PIXEL

KW - EVOLUTION

KW - GROWTH

UR - https://www.mendeley.com/catalogue/c7bb5ac9-8352-3929-9443-e3b0f1143805/

U2 - 10.3390/nano10061161

DO - 10.3390/nano10061161

M3 - Article

C2 - 32545726

VL - 10

JO - Nanomaterials

JF - Nanomaterials

SN - 2079-4991

IS - 6

M1 - 1161

ER -

ID: 26086530