Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1172
T. Cheng, H. Jin
WSD can effectively remove random noise of a raw image from very low density to ultra-high density fluorescent molecular distribution scenarios. The size of the raw image that WSD can denoise is subject to the used measurement matrix. A large raw image must be divided into blocks so that WSD denoises each block separately. Based on traditional single-molecule localization and super-resolution reconstruction scenarios, wide spectrum denoising (WSD) for blocks of different sizes was studied. The denoising ability is related to block sizes. The general trend is when the block gets larger, the denoising effect gets worse. When the block size is equal to 10, the denoising effect is the best. Using compressed sensing, only 20 raw images are needed for reconstruction. The temporal resolution is less than half a second. The spatial resolution is also greatly improved.
{"title":"Super-resolution microscopy based on wide spectrum denoising and compressed sensing","authors":"T. Cheng, H. Jin","doi":"10.18287/2412-6179-co-1172","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1172","url":null,"abstract":"WSD can effectively remove random noise of a raw image from very low density to ultra-high density fluorescent molecular distribution scenarios. The size of the raw image that WSD can denoise is subject to the used measurement matrix. A large raw image must be divided into blocks so that WSD denoises each block separately. Based on traditional single-molecule localization and super-resolution reconstruction scenarios, wide spectrum denoising (WSD) for blocks of different sizes was studied. The denoising ability is related to block sizes. The general trend is when the block gets larger, the denoising effect gets worse. When the block size is equal to 10, the denoising effect is the best. Using compressed sensing, only 20 raw images are needed for reconstruction. The temporal resolution is less than half a second. The spatial resolution is also greatly improved.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"25 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78194025","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1230
A. Yurin, G. Vishnyakov, V. Minaev
Methods for measuring the refractive index of optically transparent dielectric materials are considered. Modified methods based on the methods of minimum deviation and constant deviation are proposed and allow determining the refractive index of triangular prisms with unknown apex angles. In the proposed methods, the angles of light deviation on three faces of the prism are measured, and the refractive index of the material and the prism angles are determined from the solution of a system of equations. To implement the proposed methods, a goniometric system is used. That system was designed to measure angles between the flat surfaces of objects in manual and automated modes. Reference prism samples made of N-SF 1 optical glass, and a hollow prism filled with distilled water are studied. The proposed methods are compared and the measurement error is estimated. It is shown that the modified methods can be used for high-precision measurements of the refractive index in cases where the angles of the prism are unknown, or their measurement is associated with technical difficulties.
{"title":"Modified prism methods for measuring the refractive index of solid and liquid substances","authors":"A. Yurin, G. Vishnyakov, V. Minaev","doi":"10.18287/2412-6179-co-1230","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1230","url":null,"abstract":"Methods for measuring the refractive index of optically transparent dielectric materials are considered. Modified methods based on the methods of minimum deviation and constant deviation are proposed and allow determining the refractive index of triangular prisms with unknown apex angles. In the proposed methods, the angles of light deviation on three faces of the prism are measured, and the refractive index of the material and the prism angles are determined from the solution of a system of equations. To implement the proposed methods, a goniometric system is used. That system was designed to measure angles between the flat surfaces of objects in manual and automated modes. Reference prism samples made of N-SF 1 optical glass, and a hollow prism filled with distilled water are studied. The proposed methods are compared and the measurement error is estimated. It is shown that the modified methods can be used for high-precision measurements of the refractive index in cases where the angles of the prism are unknown, or their measurement is associated with technical difficulties.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"24 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83216900","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1232
D. Bykov, E. Bezus, L. Doskolovich
We propose a coupled-mode theory for resonant diffraction gratings with a varying period. We consider diffractive structures, in which the reciprocal lattice vector, a quantity inversely proportional to the period, varies linearly in the direction of periodicity. It is shown that optical properties of such a structure essentially depend on the period change rate. On the basis of a comparison with the results of rigorous numerical simulations using the rigorous coupled-wave analysis, high accuracy of the proposed theoretical model is demonstrated. In particular, the developed coupled-mode theory describes the broadening of the resonant peak and the appearance of secondary maxima caused by a non-zero period change rate. The obtained results can be used for the development of linear variable filters based on resonant diffraction gratings with varying parameters.
{"title":"Coupled-mode theory for resonant gratings with a varying period","authors":"D. Bykov, E. Bezus, L. Doskolovich","doi":"10.18287/2412-6179-co-1232","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1232","url":null,"abstract":"We propose a coupled-mode theory for resonant diffraction gratings with a varying period. We consider diffractive structures, in which the reciprocal lattice vector, a quantity inversely proportional to the period, varies linearly in the direction of periodicity. It is shown that optical properties of such a structure essentially depend on the period change rate. On the basis of a comparison with the results of rigorous numerical simulations using the rigorous coupled-wave analysis, high accuracy of the proposed theoretical model is demonstrated. In particular, the developed coupled-mode theory describes the broadening of the resonant peak and the appearance of secondary maxima caused by a non-zero period change rate. The obtained results can be used for the development of linear variable filters based on resonant diffraction gratings with varying parameters.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"68 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85733206","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1190
S. Chukanov
The use of traditional methods of algebraic topology to obtain information about the shape of an object is associated with the problem of forming a small amount of information, namely, Betti numbers and Euler characteristics. The central tool for topological data analysis is the persistent homology method, which summarizes the geometric and topological information in the data using persistent diagrams and barcodes. Based on persistent homology methods, topological data can be analyzed to obtain information about the shape of an object. The construction of persistent barcodes and persistent diagrams in computational topology does not allow one to construct a Hilbert space with a scalar product. The possibility of applying the methods of topological data analysis is based on mapping persistent diagrams into a Hilbert space; one of the ways of such mapping is a method of constructing a persistence landscape. It has an advantage of being reversible, so it does not lose any information and has persistence properties. The paper considers mathematical models and functions for representing persistence landscape objects based on the persistent homology method. Methods for converting persistent barcodes and persistent diagrams into persistence landscape functions are considered. Associated with persistence landscape functions is a persistence landscape kernel that forms a mapping into a Hilbert space with a dot product. A formula is proposed for determining a distance between the persistence landscapes, which allows the distance between images of objects to be found. The persistence landscape functions map persistent diagrams into a Hilbert space. Examples of determining the distance between images based on the construction of persistence landscape functions for these images are given. Representations of topological characteristics in various models of computational topology are considered. Results for one-parameter persistence modules are extended onto multi-parameter persistence modules.
{"title":"Formation of features based on computational topology methods","authors":"S. Chukanov","doi":"10.18287/2412-6179-co-1190","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1190","url":null,"abstract":"The use of traditional methods of algebraic topology to obtain information about the shape of an object is associated with the problem of forming a small amount of information, namely, Betti numbers and Euler characteristics. The central tool for topological data analysis is the persistent homology method, which summarizes the geometric and topological information in the data using persistent diagrams and barcodes. Based on persistent homology methods, topological data can be analyzed to obtain information about the shape of an object. The construction of persistent barcodes and persistent diagrams in computational topology does not allow one to construct a Hilbert space with a scalar product. The possibility of applying the methods of topological data analysis is based on mapping persistent diagrams into a Hilbert space; one of the ways of such mapping is a method of constructing a persistence landscape. It has an advantage of being reversible, so it does not lose any information and has persistence properties. The paper considers mathematical models and functions for representing persistence landscape objects based on the persistent homology method. Methods for converting persistent barcodes and persistent diagrams into persistence landscape functions are considered. Associated with persistence landscape functions is a persistence landscape kernel that forms a mapping into a Hilbert space with a dot product. A formula is proposed for determining a distance between the persistence landscapes, which allows the distance between images of objects to be found. The persistence landscape functions map persistent diagrams into a Hilbert space. Examples of determining the distance between images based on the construction of persistence landscape functions for these images are given. Representations of topological characteristics in various models of computational topology are considered. Results for one-parameter persistence modules are extended onto multi-parameter persistence modules.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"31 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81099383","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1209
M. Kataev, E. Kartashov, V. D. Avdeenko, Evgeny Yurievich, Kartashov
With road defects being a key factor in traffic accidents, driver safety, vehicle condition and travel speed, they need to be promptly repaired. The defects include cracks, ruts and potholes on the road surface, and if not repaired in due time, they grow in size very quickly. There are various methods that are used to detect road defects, one of which is a computer vision method. Typically, digital cameras are installed on cars and then the resulting set of images from a given section of the road is processed. This article proposes a technique for obtaining images using unmanned aerial vehicles, which provides the required amount of data to assess the road condition on lengthy road sections. A technique for identifying road cracks and estimating parameters defined in the road traffic regulations is proposed. As a result of the studies, real images are obtained and processed using the method proposed herein, showing high performance and accuracy and indicating the possibility of its future practical uses.
{"title":"Method for detecting road defects using images obtained from unmanned aerial vehicles","authors":"M. Kataev, E. Kartashov, V. D. Avdeenko, Evgeny Yurievich, Kartashov","doi":"10.18287/2412-6179-co-1209","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1209","url":null,"abstract":"With road defects being a key factor in traffic accidents, driver safety, vehicle condition and travel speed, they need to be promptly repaired. The defects include cracks, ruts and potholes on the road surface, and if not repaired in due time, they grow in size very quickly. There are various methods that are used to detect road defects, one of which is a computer vision method. Typically, digital cameras are installed on cars and then the resulting set of images from a given section of the road is processed. This article proposes a technique for obtaining images using unmanned aerial vehicles, which provides the required amount of data to assess the road condition on lengthy road sections. A technique for identifying road cracks and estimating parameters defined in the road traffic regulations is proposed. As a result of the studies, real images are obtained and processed using the method proposed herein, showing high performance and accuracy and indicating the possibility of its future practical uses.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"34 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91238907","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1204
D. Prozorov, A. Zemtsov
The article analyzes the possibility of using a Siamese convolutional neural network to solve the problem of vascular authentication on an embedded hardware platform with limited computing resources (Orange Pi One). The authors give a brief review of modern methods for calculating image feature vectors used in the tasks of classifying, comparing or searching for images by content: based on variational series (histograms), local descriptors, singular point descriptors, descriptors based on hash functions, neural network descriptors. They suggest using the architecture of a biometric authentication system (BAS) based on images of palms in the visible and near-IR spectra based on a Siamese convolutional neural network. The developed software solution allows using the Siamese neural network in the "full network" (both symmetrical channels of the neural network are used) and "half of the neural network" (only one channel is used) modes to reduce the time for comparing biometric data vectors - images of the palms of registered BAS users. The authors demonstrate advantages of the neural network features: universality, scalability and competitiveness, including on embedded hardware and software solutions with limited computing resources without graphics accelerators. The studies have shown that using the Siamese neural network, the "overall accuracy" of palm image classification can be improved from 0.929 to 0.968 when compared with the image vectorization method based on a perceptual hash, while showing a comparable authentication time for individuals registered in BAS. In the experiments, the authors use a database of 2,000 images for 400 people.
本文分析了在计算资源有限的嵌入式硬件平台(Orange Pi One)上使用连体卷积神经网络解决血管认证问题的可能性。作者简要回顾了用于图像分类、比较或按内容搜索任务的图像特征向量计算的现代方法:基于变分序列(直方图)、局部描述符、奇点描述符、基于哈希函数的描述符、神经网络描述符。他们建议使用基于暹罗卷积神经网络的手掌可见光和近红外光谱图像的生物识别认证系统(BAS)架构。开发的软件解决方案允许在“全网络”(使用神经网络的两个对称通道)和“半神经网络”(仅使用一个通道)模式下使用Siamese神经网络，以减少比较生物特征数据向量(注册BAS用户的手掌图像)的时间。作者展示了神经网络特征的优势:通用性、可扩展性和竞争力，包括在没有图形加速器的有限计算资源的嵌入式硬件和软件解决方案上。研究表明，使用Siamese神经网络，与基于感知哈希的图像矢量化方法相比，掌纹图像分类的“整体准确率”可以从0.929提高到0.968，同时对在BAS中注册的个体具有相当的认证时间。在实验中，作者使用了一个包含400人的2000张图像的数据库。
{"title":"Using a lightweight Siamese neural network for generating a feature vector in a vascular authentication system","authors":"D. Prozorov, A. Zemtsov","doi":"10.18287/2412-6179-co-1204","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1204","url":null,"abstract":"The article analyzes the possibility of using a Siamese convolutional neural network to solve the problem of vascular authentication on an embedded hardware platform with limited computing resources (Orange Pi One). The authors give a brief review of modern methods for calculating image feature vectors used in the tasks of classifying, comparing or searching for images by content: based on variational series (histograms), local descriptors, singular point descriptors, descriptors based on hash functions, neural network descriptors. They suggest using the architecture of a biometric authentication system (BAS) based on images of palms in the visible and near-IR spectra based on a Siamese convolutional neural network. The developed software solution allows using the Siamese neural network in the \"full network\" (both symmetrical channels of the neural network are used) and \"half of the neural network\" (only one channel is used) modes to reduce the time for comparing biometric data vectors - images of the palms of registered BAS users. The authors demonstrate advantages of the neural network features: universality, scalability and competitiveness, including on embedded hardware and software solutions with limited computing resources without graphics accelerators. The studies have shown that using the Siamese neural network, the \"overall accuracy\" of palm image classification can be improved from 0.929 to 0.968 when compared with the image vectorization method based on a perceptual hash, while showing a comparable authentication time for individuals registered in BAS. In the experiments, the authors use a database of 2,000 images for 400 people.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"27 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78003635","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1228
E. Kozlova, A. Savelyeva, A. Kovalev, V. Kotlyar
It is shown that a squared Laguerre-Gaussian (LG) vortex beam is Fourier-invariant and retains its structure at the focus of a spherical lens. In the Fresnel diffraction zone, such a beam is transformed into superposition of conventional LG beams, the number of which is equal to the number of rings in the squared LG beam. If there is only one ring, then the beam is structurally stable. A more general beam, which is a “product” of two LG beams, is also considered. Such a beam will be Fourier-invariant if the number of rings in two LG beams in the “product” is the same. The considered beams complement the well-known family of LG beams, which are intensively studied as they remain stable during their propagation in turbulent media.
{"title":"A Fourier-invariant squared Laguerre-Gaussian vortex beam","authors":"E. Kozlova, A. Savelyeva, A. Kovalev, V. Kotlyar","doi":"10.18287/2412-6179-co-1228","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1228","url":null,"abstract":"It is shown that a squared Laguerre-Gaussian (LG) vortex beam is Fourier-invariant and retains its structure at the focus of a spherical lens. In the Fresnel diffraction zone, such a beam is transformed into superposition of conventional LG beams, the number of which is equal to the number of rings in the squared LG beam. If there is only one ring, then the beam is structurally stable. A more general beam, which is a “product” of two LG beams, is also considered. Such a beam will be Fourier-invariant if the number of rings in two LG beams in the “product” is the same. The considered beams complement the well-known family of LG beams, which are intensively studied as they remain stable during their propagation in turbulent media.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"105 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77348863","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1243
A. Volyar, E. Abramochkin, Y. Akimova, M. Bretsko
The article investigated conditions for shaping astigmatic-invariant structured Laguerre-Gaussian (LG) beams in the case of simple astigmatism. We have theoretically and experimentally confirmed that the conditions of astigmatic invariance are that the q-phase parameter of the structured LG (sLG) beam is equal to the arctangent of the ratio of the Rayleigh length z0 to the focal length fcyl of a cylindrical lens for a single amplitude parameter of epsilon;=1. For the rest amplitude parameter values, epsilon;NotEqual;1, the astigmatic invariance condition is set by the equality of the orbital angular momenta (OAM) of structured sLG and astigmatic-invariant sLG (asLG) beams. We have also found sharp spikes and dips of the OAM in astigmatic asLG beams in the region where OAM turns into zero. The height and depth of these spikes and dips significantly exceed the maximum and minimum OAM values in the conventional structured sLG beams. It has been shown that the occurrence of spikes and dips of the OAM is caused by a radical restructuring of the LG mode spectra in the form of their strong ordering. Theoretical calculations, accompanied by a computer simulation, and an experiment agree well with each other.
{"title":"Huge spikes and dips of the orbital angular momentum in structured Laguerre-Gaussian beams resistant to simple astigmatism","authors":"A. Volyar, E. Abramochkin, Y. Akimova, M. Bretsko","doi":"10.18287/2412-6179-co-1243","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1243","url":null,"abstract":"The article investigated conditions for shaping astigmatic-invariant structured Laguerre-Gaussian (LG) beams in the case of simple astigmatism. We have theoretically and experimentally confirmed that the conditions of astigmatic invariance are that the q-phase parameter of the structured LG (sLG) beam is equal to the arctangent of the ratio of the Rayleigh length z0 to the focal length fcyl of a cylindrical lens for a single amplitude parameter of epsilon;=1. For the rest amplitude parameter values, epsilon;NotEqual;1, the astigmatic invariance condition is set by the equality of the orbital angular momenta (OAM) of structured sLG and astigmatic-invariant sLG (asLG) beams. We have also found sharp spikes and dips of the OAM in astigmatic asLG beams in the region where OAM turns into zero. The height and depth of these spikes and dips significantly exceed the maximum and minimum OAM values in the conventional structured sLG beams. It has been shown that the occurrence of spikes and dips of the OAM is caused by a radical restructuring of the LG mode spectra in the form of their strong ordering. Theoretical calculations, accompanied by a computer simulation, and an experiment agree well with each other.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"9 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90357966","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1219
I.M. Sizova
Profile transformation of a uniform monochromatic beam with a flat wavefront after the propagation through an apodizer comprising a square serrated aperture and a spatial filter is theoretically considered at the apodizer exit and behind it. Main differences from the case of the circular apodizer geometry are shown. Roles of the square serrated aperture angles, the phase in Fresnel diffraction, the shape of the aperture serrations and serration height-to-period ratio are analyzed.
{"title":"Laser beam shaping in an apodizer with a rectangular serrated aperture","authors":"I.M. Sizova","doi":"10.18287/2412-6179-co-1219","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1219","url":null,"abstract":"Profile transformation of a uniform monochromatic beam with a flat wavefront after the propagation through an apodizer comprising a square serrated aperture and a spatial filter is theoretically considered at the apodizer exit and behind it. Main differences from the case of the circular apodizer geometry are shown. Roles of the square serrated aperture angles, the phase in Fresnel diffraction, the shape of the aperture serrations and serration height-to-period ratio are analyzed.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"43 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87461255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.18287/2412-6179-co-1203
A. Orlov, E. S. Abramova
In this paper, the relevance of developing methods and algorithms for neural network incremental learning is shown. Families of incremental learning techniques are presented. A possibility of using the extreme learning machine for incremental learning is assessed. Experiments show that the extreme learning machine is suitable for incremental learning, but as the number of training examples increases, the neural network becomes unsuitable for further learning. To solve this problem, we propose a neural network incremental learning algorithm that alternately uses the extreme learning machine to correct the only output layer network weights (operation mode) and the backpropagation method (deep learning) to correct all network weights (sleep mode). During the operation mode, the neural network is assumed to produce results or learn from new tasks, optimizing its weights in the sleep mode. The proposed algorithm features the ability for real-time adaption to changing external conditions in the operation mode. The effectiveness of the proposed algorithm is shown by an example of solving the approximation problem. Approximation results after each step of the algorithm are presented. A comparison of the mean square error values when using the extreme learning machine for incremental learning and the developed algorithm of neural network alternate incremental learning is made.
{"title":"Development and research of a neural network alternate incremental learning algorithm","authors":"A. Orlov, E. S. Abramova","doi":"10.18287/2412-6179-co-1203","DOIUrl":"https://doi.org/10.18287/2412-6179-co-1203","url":null,"abstract":"In this paper, the relevance of developing methods and algorithms for neural network incremental learning is shown. Families of incremental learning techniques are presented. A possibility of using the extreme learning machine for incremental learning is assessed. Experiments show that the extreme learning machine is suitable for incremental learning, but as the number of training examples increases, the neural network becomes unsuitable for further learning. To solve this problem, we propose a neural network incremental learning algorithm that alternately uses the extreme learning machine to correct the only output layer network weights (operation mode) and the backpropagation method (deep learning) to correct all network weights (sleep mode). During the operation mode, the neural network is assumed to produce results or learn from new tasks, optimizing its weights in the sleep mode. The proposed algorithm features the ability for real-time adaption to changing external conditions in the operation mode. The effectiveness of the proposed algorithm is shown by an example of solving the approximation problem. Approximation results after each step of the algorithm are presented. A comparison of the mean square error values when using the extreme learning machine for incremental learning and the developed algorithm of neural network alternate incremental learning is made.","PeriodicalId":46692,"journal":{"name":"Computer Optics","volume":"38 1","pages":""},"PeriodicalIF":1.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84938839","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：