• Journal of Korea Multimedia Society
• /
• v.24 no.11
• /
• pp.1552-1559
• /
• 2021

Convolution neural networks (CNNs) show notable performance in image processing and are used as representative core models. CNNs extract and learn features from large amounts of train dataset. In general, it has a structure in which a convolution layer and a fully connected layer are stacked. The core of CNN is the convolution layer. The size of the kernel used for feature extraction and the number that affect the depth of the feature map determine the amount of weight parameters of the CNN that can be learned. These parameters are the main causes of increasing the computational complexity and memory usage of the entire neural network. The most computationally expensive components in CNNs are fully connected and spatial convolution computations. In this paper, we propose a Fourier Convolution Neural Network that performs the operation of the convolution layer in the Fourier domain. We work on modifying and improving the amount of computation by applying the fast fourier transform method. Using the MNIST dataset, the performance was similar to that of the general CNN in terms of accuracy. In terms of operation speed, 7.2% faster operation speed was achieved. An average of 19% faster speed was achieved in experiments using 1024x1024 images and various sizes of kernels.

• Journal of Korean Association for Spatial Structures
• /
• v.4 no.2 s.12
• /
• pp.81-88
• /
• 2004

The dynamic instability of snapping phenomena has been studied by many researchers. Few papers deal with dynamic buckling under loads with periodic characteristics, and the behavior under periodic excitations is expected to be different from behavior under STEP excitations. We investigate the fundamental mechanisms of the dynamic instability when the sinusoidally shaped arch structures are subjected to sinusoidally distributed excitations with pin-ends. The mechanisms of dynamic indirect snapping of shallow arches are especially investigated under not only STEP function excitations but also under sinusoidal harmonic excitations, applied i the up-and-down direction. The dynamic nonlinear responses are obtained by the numerical integration of the geometrically nonlinear equation of motion. And using this analyze characteristics of the dynamic instability through the running response spectrum by FFT(Fast Fourier Transform).

• Journal of the Society of Naval Architects of Korea
• /
• v.42 no.3
• /
• pp.212-219
• /
• 2005

Real sea waves in a towing wave basin have been generated using random periodic motion of the segmented wave makers and the wave reflections of sidewalls. Theoretically, the real sea waves can be described by the superposition of many random oblique waves. This paper introduces numerical real sea wave generation in a rectangular wave basin using spectral method that uses a superposition of orthogonal functions which have to satisfy the Laplace equation. Oblique regular waves, long crested irregular waves and real sea waves were simulated and met the requirement of sidewall wave reflection and wave absorption. MLM (Maximum Likelihood Method) and Spatial Fourier Transform were used in order to obtain propagated wave direction characteristics. The estimated results proved the usefulness of the method and the performances showed reasonable directional patterns comparing with generating patterns.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2019.05a
• /
• pp.439-441
• /
• 2019

Deep learning has become a powerful and robust algorithm in Artificial Intelligence. One of the most impressive forms of Deep learning tools is that of the Convolutional Neural Networks (CNN). CNN is a state-of-the-art solution for object recognition. For instance when we utilize CNN with MNIST handwritten digital dataset, mostly the result is well. Because, in MNIST dataset, all digits are centralized. Unfortunately, the real world is different from our imagination. If digits are shifted from the center, it becomes a big issue for CNN to recognize and provide result like before. To solve that issue, we have created frequency images from spatial images by a Fast Fourier Transform (FFT).

• Journal of Korea Multimedia Society
• /
• v.19 no.10
• /
• pp.1775-1781
• /
• 2016

Generally, in the underwater target recognition, feature vectors are extracted from the target signal utilizing spatial information according to target shape/material characteristics. In addition, various signal processing techniques have been studied to extract feature vectors which are less sensitive to the location of the receiver. In this paper, we synthesized active echo signals using 3-dimensional highlight distribution. Then, Fractional Fourier transform was applied to echo signals to extract signal features. For the performance verification, classification experiments were performed using backpropagation and probabilistic neural network classifiers based on single aspect and multi-aspect method. As a result, we obtained a better recognition result using proposed feature extraction and multi-aspect based method.

• Journal of the Korean Society of Safety
• /
• v.13 no.4
• /
• pp.256-263
• /
• 1998

In this paper, we propose a volume holographic fingerprint recognition system based on optical correlator for personal identification. Optical correlator has high speed and parallel processing characteristics of optics. Matched filters are recorded into a volume hologram that can store data with high density, transfer them with high speed, and select a randomly chosen data element. The multiple reference images of database are prerecorded in a photorefractive crystal in the form of Fourier transform images, simply by passing the image displayed in a spatial light modulator through a Fourier transform lens. The angular multiplexing method for multiple holograms of database can be achieved by rotating the crystal by use of a step motor. Experimental results show that the proposed system can be used for the security verification system.

• Journal of applied mathematics & informatics
• /
• v.19 no.1_2
• /
• pp.179-190
• /
• 2005

We deal with the Cauchy problem for the space-time fractional diffusion equation, which is obtained from standard diffusion equation by replacing the second-order space derivative with a Caputo (or Riemann-Liouville) derivative of order ${\beta}{\in}$ (0, 2] and the first-order time derivative with Caputo derivative of order ${\beta}{\in}$ (0, 1]. The fundamental solution (Green function) for the Cauchy problem is investigated with respect to its scaling and similarity properties, starting from its Fourier-Laplace representation. We derive explicit expression of the Green function. The Green function also can be interpreted as a spatial probability density function evolving in time. We further explain the similarity property by discussing the scale-invariance of the space-time fractional diffusion equation.

• Journal of applied mathematics & informatics
• /
• v.18 no.1_2
• /
• pp.339-350
• /
• 2005

A space-time fractional advection-dispersion equation (ADE) is a generalization of the classical ADE in which the first-order time derivative is replaced with Caputo derivative of order $\alpha{\in}(0,1]$, and the second-order space derivative is replaced with a Riesz-Feller derivative of order $\beta{\in}0,2]$. We derive the solution of its Cauchy problem in terms of the Green functions and the representations of the Green function by applying its Fourier-Laplace transforms. The Green function also can be interpreted as a spatial probability density function (pdf) evolving in time. We do the same on another kind of space-time fractional advection-dispersion equation whose space and time derivatives both replacing with Caputo derivatives.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2011.04a
• /
• pp.250-257
• /
• 2011

The power input to an infinite cylindrical shell excited by a point force is investigated. The circumferential direction and axial direction of the cylindrical shell is assumed as a two-dimensional unbounded medium, and the point force is replaced as a periodic array of imaginary sources. The spatial Fourier transform is taken from the equation of motion of the cylindrical shell, which is derived from the static model of Donell-Mushtari-Vlasov. The inverse Fourier transform is taken to derive the vibration responses. Mobility from out-of-plane forces and in-plane forces are derived from the obtained vibration responses. The theory is applied to a cylindrical shell excited by a normal direction of point force.

• ETRI Journal
• /
• v.16 no.1
• /
• pp.1-15
• /
• 1994

A computationally-efficient approach to the calculation of the transient field of an acoustic radiator was developed. With this approach, a planar or curved source, radiating either continuous or pulsed waves, is divided into a finite number of shifted and/or rotated versions of an incremental source such that the Fraunhofer approximation holds at each field point. The acoustic field from the incremental source is given by a 2-D spatial Fourier transform. The diffraction transfer function of the entire source can be expressed as a sum of Fraunhofer diffraction pattern of the incremental sources with the appropriate coordinate transformations for the particular geometry of the radiator. For a given spectrum of radiator velocity, the transient field can be computed directly in the frequency domain using the diffraction transfer function. To determine the accuracy of the proposed approach, the impulse response was derived using the inverse Fourier transform. The results obtained agree well with published data obtained using the impulse response approach. The computational efficiency of the proposed method compares favorably to those of the point source method and the impulse response approach.