RU99120927A - SIGNAL PROCESSING METHOD - Google Patents

SIGNAL PROCESSING METHOD

Info

Publication number
RU99120927A
RU99120927A RU99120927/09A RU99120927A RU99120927A RU 99120927 A RU99120927 A RU 99120927A RU 99120927/09 A RU99120927/09 A RU 99120927/09A RU 99120927 A RU99120927 A RU 99120927A RU 99120927 A RU99120927 A RU 99120927A
Authority
RU
Russia
Prior art keywords
signal
neural network
network
information
input information
Prior art date
Application number
RU99120927/09A
Other languages
Russian (ru)
Other versions
RU2189078C2 (en
Inventor
Игорь Викторович Денисов
Олег Тимурович Каменев
Юрий Николаевич Кульчин
Олег Викторович Кириченко
Original Assignee
Дальневосточный государственный технический университет
Filing date
Publication date
Application filed by Дальневосточный государственный технический университет filed Critical Дальневосточный государственный технический университет
Priority to RU99120927A priority Critical patent/RU2189078C2/en
Priority claimed from RU99120927A external-priority patent/RU2189078C2/en
Publication of RU99120927A publication Critical patent/RU99120927A/en
Application granted granted Critical
Publication of RU2189078C2 publication Critical patent/RU2189078C2/en

Links

Claims (1)

Способ обработки сигналов, включающий использование вычислительного комплекса, содержащего нейронную сеть, матрицу связей нейронов в сети, реализующую необходимое функциональное преобразование входного информационного сигнала, канал подвода этого сигнала и канал отвода выходного сигнала, предусматривающий распознавание сигнала с выработкой выходного управляющего сигнала вычислительного комплекса, отличающийся тем, что в качестве входного информационного сигнала используют оптический сигнал, который делят на два информационных потока, при этом первый из них используют для восстановления исходной информации и последующей идентификации образа, несомого входным информационным сигналом, а второй используют для реализации необходимых функциональных преобразований этого сигнала, для чего первый информационный поток последовательно пропускают через оптическую матрицу связей однопроходной нейронной сети типа персептрон, восстанавливая его, после чего восстановленный оптический сигнал подают на вход многопроходной классифицирующей нейронной сети, например, типа сети Хопфилда, причем выходной сигнал этой сети используют для управления выбором матрицы связей однопроходной нейронной сети типа персептрон, реализующей необходимое функциональное преобразование второго информационного потока входного информационного сигнала.A method of processing signals, including the use of a computer complex containing a neural network, a matrix of neuron connections in the network that implements the necessary functional transformation of the input information signal, a channel for supplying this signal and a channel for extracting the output signal, providing for signal recognition with the generation of an output control signal for the computer complex, characterized in that as an input information signal using an optical signal, which is divided into two information p outflow, while the first of them is used to restore the initial information and subsequent identification of the image carried by the input information signal, and the second is used to implement the necessary functional transformations of this signal, for which the first information stream is sequentially passed through the optical matrix of communications of a single-pass neural network such as perceptron, restoring it, after which the restored optical signal is fed to the input of a multi-pass classifying neural network, for example, of the type Hopfield network, and the output signal of this network is used to control the selection of the matrix of communications of a single-pass neural network such as perceptron, which implements the necessary functional transformation of the second information stream of the input information signal.
RU99120927A 1999-10-01 1999-10-01 Signal processing method RU2189078C2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
RU99120927A RU2189078C2 (en) 1999-10-01 1999-10-01 Signal processing method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
RU99120927A RU2189078C2 (en) 1999-10-01 1999-10-01 Signal processing method

Publications (2)

Publication Number Publication Date
RU99120927A true RU99120927A (en) 2001-08-27
RU2189078C2 RU2189078C2 (en) 2002-09-10

Family

ID=20225490

Family Applications (1)

Application Number Title Priority Date Filing Date
RU99120927A RU2189078C2 (en) 1999-10-01 1999-10-01 Signal processing method

Country Status (1)

Country Link
RU (1) RU2189078C2 (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109211122B (en) * 2018-10-30 2020-05-15 清华大学 Ultra-precise displacement measurement system and method based on optical neural network
WO2022056422A1 (en) * 2020-09-14 2022-03-17 The Regents Of The University Of California Ensemble learning of diffractive neural networks

Similar Documents

Publication Publication Date Title
US20240028896A1 (en) Method and system for activity classification
Pasupa et al. A comparison between shallow and deep architecture classifiers on small dataset
Van Quang et al. CapsuleNet for micro-expression recognition
Reddy et al. Spontaneous facial micro-expression recognition using 3D spatiotemporal convolutional neural networks
Rahman et al. An improved attention for visual question answering
DE60128029T2 (en) Apparatus and method for detecting or recognizing patterns by means of a plurality of feature detection elements
Wang et al. Eye recognition with mixed convolutional and residual network (MiCoRe-Net)
Lee et al. Integrated segmentation and recognition of handwritten numerals with cascade neural network
Wangchuk et al. Real-time Bhutanese sign language digits recognition system using convolutional neural network
Mani et al. Application of artificial neural network model for optical character recognition
US7512571B2 (en) Associative memory device and method based on wave propagation
Xu et al. Recurrent convolutional neural network for video classification
Serrano-Gotarredona et al. ConvNets experiments on SpiNNaker
Chen et al. Learning capsules for vehicle logo recognition
GB2245401A (en) Neural network signal processor
Musaev et al. Accelerated training for convolutional neural networks
RU99120927A (en) SIGNAL PROCESSING METHOD
Hossain et al. Sign language digit recognition using different convolutional neural network model
Rabbi et al. Kdanet: Handwritten character recognition for Bangla language using deep learning
Malik et al. Handwritten character recognition using wavelet transform and hopfield network
Espinel et al. Face gesture recognition using deep-learning models
Khadhouri et al. Compound effects of top-down and bottom-up influences on visual attention during action recognition
JPH0581227A (en) Neuron system network signal processor and method of processing signal
Thorat et al. Modulation of early visual processing alleviates capacity limits in solving multiple tasks
Marais et al. Spatiotemporal Convolutions and Video Vision Transformers for Signer-Independent Sign Language Recognition