Improved Measurement of Thin Film Thickness in Spectroscopic Reflectometer Using Convolutional Neural Networks

Kim, Min-Gab

doi:10.1007/s12541-019-00260-4

Cited by 17 publications

(14 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…More importantly, in some previous works, when the difference between the ANN algorithm and the model-based algorithm was found to be relatively large, the output of the ANN algorithm was used as the initial value of the model-based algorithm to reduce the difference 15 , 18 . This allows the iteration steps of the model-based algorithm to be reduced.…”

Section: Resultsmentioning

confidence: 99%

“…Even if the difference is small enough, it doesn’t mean that the ANN algorithm works accurately because both analysis methods can be incorrect. Therefore, because of insufficient verification on the effectiveness of the ANN algorithms, the results obtained with an ANN algorithm should only be used as initial values in a model-based analysis to reduce the difference with less iteration 15 , 18 . Despite of the practical difficulty regarding analysis reliability, no methods beyond the simple comparison have been proposed and adopted for verifying the ANN algorithms.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A novel method to design and evaluate artificial neural network for thin film thickness measurement traceable to the length standard

Lee

Jin

2022

Sci Rep

View full text Add to dashboard Cite

The artificial neural networks (ANNs) have been often used for thin-film thickness measurement, whose performance evaluations were only conducted at the level of simple comparisons with the existing analysis methods. However, it is not an easy and simple way to verify the reliability of an ANN based on international length standards. In this article, we propose for the first time a method by which to design and evaluate an ANN for determining the thickness of the thin film with international standards. The original achievements of this work are to choose parameters of the ANN reasonably and to evaluate the training instead of a simple comparison with conventional methods. To do this, ANNs were built in 12 different cases, and then trained using theoretical spectra. The experimental spectra of the certified reference materials (CRMs) used here served as the validation data of each trained ANN, with the output then compared with a certified value. When both values agree with each other within an expanded uncertainty of the CRMs, the ANN is considered to be reliable. We expect that the proposed method can be useful for evaluating the reliability of ANN in the future.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A novel method to design and evaluate artificial neural network for thin film thickness measurement traceable to the length standard

Lee

Jin

2022

Sci Rep

View full text Add to dashboard Cite

show abstract

“…[12][13][14] The capabilities of deep neural networks result from their ability to learn complex patterns by generalizing large quantities of data through the training of a large number of internal weight parameters. 15,16 These networks have shown the remarkable ability to solve the ID problem accurately and efficiently in specific nanophotonic systems such as thin-films, 10,[17][18][19][20] 2D metasurfaces, [21][22][23][24][25][26][27] and core-shell nanoparticles, 8,28 within a limited parameter space. Furthermore, neural network models have been implemented to replace the typical forward electromagnetic methods used to simulate optical systems.…”

Section: Introductionmentioning

confidence: 99%

General Inverse Design of Thin-Film Metamaterials With Convolutional Neural Networks

Lininger,

Hinczewski,

Strangi

2021

Preprint

View full text Add to dashboard Cite

The design of metamaterials which support unique optical responses is the basis for most thin-film nanophotonics applications. In practice this inverse design problem can be difficult to solve systematically due to the large design parameter space associated with general multi-layered systems. We apply convolutional neural networks, a subset of deep machine learning, as a tool to solve this inverse design problem for metamaterials composed of stacks of thin films. We demonstrate the remarkable ability of neural networks to probe the large global design space (up to 10 12 possible parameter combinations) and resolve all relationships between metamaterial structure and corresponding ellipsometric and reflectance / transmittance spectra. The applicability of the approach is further expanded to include the inverse design of synthetic engineered spectra in general design scenarios. Furthermore, this approach is compared

show abstract

“…Figure 2 shows the description of several common combined creep models. This process is called feedforward operation [16]. In the last layer, the target task is formalized; that is, classification or regression is performed in the target function.…”

Section: Introductionmentioning

confidence: 99%

Image Classification Algorithm Based on Big Data and Multilabel Learning of Improved Convolutional Neural Network

Chang

Cui

2021

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

More and more image materials are used in various industries these days. Therefore, how to collect useful images from a large set has become an urgent priority. Convolutional neural networks (CNN) have achieved good results in certain image classification tasks, but there are still problems such as poor classification ability, low accuracy, and slow convergence speed. This article mainly introduces the image classification algorithm (ICA) research based on the multilabel learning of the improved convolutional neural network and some improvement ideas for the research of the ICA based on the multilabel learning of the convolutional neural network. This paper proposes an ICA research method based on multilabel learning of improved convolutional neural networks, including the image classification process, convolutional network algorithm, and multilabel learning algorithm. The conclusions show that the average maximum classification accuracy of the improved CNN in this paper is 90.63%, and the performance is better, which is beneficial to improving the efficiency of image classification. The improved CNN network structure has reached the highest accuracy rate of 91.47% on the CIFAR-10 data set, which is much higher than the traditional CNN algorithm.

show abstract

Improved Measurement of Thin Film Thickness in Spectroscopic Reflectometer Using Convolutional Neural Networks

Cited by 17 publications

References 11 publications

A novel method to design and evaluate artificial neural network for thin film thickness measurement traceable to the length standard

A novel method to design and evaluate artificial neural network for thin film thickness measurement traceable to the length standard

General Inverse Design of Thin-Film Metamaterials With Convolutional Neural Networks

Image Classification Algorithm Based on Big Data and Multilabel Learning of Improved Convolutional Neural Network

Contact Info

Product

Resources

About