• Journal of the Korea Society of Computer and Information
• /
• v.27 no.2
• /
• pp.33-42
• /
• 2022

In this paper, we proposed a deep learning model based on multi-scale dilated convolution feature fusion for the segmentation of black ice region in road image to send black ice warning to drivers in real time. In the proposed multi-scale dilated convolution feature fusion network, different dilated ratio convolutions are connected in parallel in the encoder blocks, and different dilated ratios are used in different resolution feature maps, and multi-layer feature information are fused together. The multi-scale dilated convolution feature fusion improves the performance by diversifying and expending the receptive field of the network and by preserving detailed space information and enhancing the effectiveness of diated convolutions. The performance of the proposed network model was gradually improved with the increase of the number of dilated convolution branch. The mIoU value of the proposed method is 96.46%, which was higher than the existing networks such as U-Net, FCN, PSPNet, ENet, LinkNet. The parameter was 1,858K, which was 6 times smaller than the existing LinkNet model. From the experimental results of Jetson Nano, the FPS of the proposed method was 3.63, which can realize segmentation of black ice field in real time.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.9
• /
• pp.3598-3614
• /
• 2020

With the increase of motor vehicles and tourism demand, some traffic problems gradually appear, such as traffic congestion, safety accidents and insufficient allocation of traffic resources. Facing these challenges, a model of Spatio-Temporal Dilated Convolutional Network (STDGCN) is proposed for assistance of extracting highly nonlinear and complex characteristics to accurately predict the future traffic flow. In particular, we model the traffic as undirected graphs, on which graph convolutions are built to extract spatial feature informations. Furthermore, a dilated convolution is deployed into graph convolution for capturing multi-scale contextual messages. The proposed STDGCN integrates the dilated convolution into the graph convolution, which realizes the extraction of the spatial and temporal characteristics of traffic flow data, as well as features of road occupancy. To observe the performance of the proposed model, we compare with it with four rivals. We also employ four indicators for evaluation. The experimental results show STDGCN's effectiveness. The prediction accuracy is improved by 17% in comparison with the traditional prediction methods on various real-world traffic datasets.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2007.11a
• /
• pp.1037-1041
• /
• 2007

With the advent of sound field simulator, many sound fields have been reproduced by obtaining the impulse responses of specific acoustic spaces like famous concert hall, opera house. This sound field reproduction has been done by the linear convolution operation between the sound input signal and the impulse response of certain acoustic space. However, the conventional finite impulse response based linear convolution operation always makes real-time implementation of sound field generator impossible due to the large amount of computational burden. This paper introduces the fast convolution method using perceptual redundancy in the processed signals, input audio signal and room impulse response. Temporal and spectral psycho-acoustic filters considering masking effects are implemented in the proposed convolution structure. It reduces the computational burden of convolution methods for realtime implementation of a sound field generator. The conventional convolutions are compared with the proposed one in views of computational burden and sound quality. In the proposed method, a considerable reduction in the computational burden was realized with acceptable changes in sound quality.

• Bulletin of the Korean Mathematical Society
• /
• v.52 no.1
• /
• pp.105-123
• /
• 2015

Let $f_{\beta}=h_{\beta}+\bar{g}_{\beta}$ and $F_a=H_a+\bar{G}_a$ be harmonic mappings obtained by shearing of analytic mappings $h_{\beta}+g_{\beta}=1/(2isin{\beta})log\((1+ze^{i{\beta}})/(1+ze^{-i{\beta}})\)$, 0 < ${\beta}$ < ${\pi}$ and $H_a+G_a=z/(1-z)$, respectively. Kumar et al. [7] conjectured that if ${\omega}(z)=e^{i{\theta}}z^n({\theta}{\in}\mathbb{R},n{\in}\mathbb{N})$ and ${\omega}_a(z)=(a-z)/(1-az)$, $a{\in}(-1,1)$ are dilatations of $f_{\beta}$ and $F_a$, respectively, then $F_a\tilde{\ast}f_{\beta}{\in}S^0_H$ and is convex in the direction of the real axis, provided $a{\in}[(n-2)/(n+2),1)$. They claimed to have verified the result for n = 1, 2, 3 and 4 only. In the present paper, we settle the above conjecture, in the affirmative, for ${\beta}={\pi}/2$ and for all $n{\in}\mathbb{N}$.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.18 no.2
• /
• pp.231-237
• /
• 2008

With the advent of sound field simulator, many sound fields have been reproduced by obtaining the impulse responses of specific acoustic spaces like famous concert hall, opera house. This sound field reproduction has been done by the linear convolution operation between the sound input signal and the impulse response of certain acoustic space. However, the conventional finite impulse response based linear convolution operation always makes real-time implementation of sound field generator impossible due to the large amount of computational burden. This paper introduces the fast convolution method using perceptual redundancy in the processed signals, input audio signal and room impulse response. Temporal and spectral real-time masking blocks are implemented in the proposed convolution structure. It reduces the computational burden of convolution methods for real-time implementation of a sound field generator. The conventional convolutions are compared with the proposed one in views of computational burden and sound quality. In the proposed method, a considerable reduction in the computational burden was realized with acceptable changes in sound quality.

• Journal of Advanced Marine Engineering and Technology
• /
• v.23 no.5
• /
• pp.702-710
• /
• 1999

U-shaped bellows are usually used to piping system pressure sensor and controller for refriger-ator. Bellows subjected to internal pressure are designed for the purpose of absorbing deformation. Internal pressure on the convolution sidewall and end collar will be applied to an axial load tend-ing to push the collar away from the convolutions. To find out deformation behavior of bellow sub-jected to internal pressure the axisymmetric shell theory using the finite element method is adopted in this paper. U-shaped bellows can be idealized by series of conical frustum-shaped ele-ments because it is axisymmetric shell structure. The displacements of nodal points due to small increment of force are calculated by the finite element method and the calculated nodal displace-ments are added to r-z cylindrical coordinates of nodal points. The new stiffness matrix of the sys-tem using the new coordinates of nodal points is adopted to calculate the another increments of nodal displacement that is the step by step method is used in this paper. The force required to deflect bellows axially is a function of the dimensions of the bellows and the materials from which they are made. Spring constant is analyzed according to the changing geometric factors of U-shaped bellows. The FEM results were agreed with experiment. Using developed FORTRAN PROGRAM the internal pressure vs. deflection characteristics of a particu-lar bellows can be predicted by input of a few factors.

• International Journal of Computer Science & Network Security
• /
• v.21 no.9
• /
• pp.203-211
• /
• 2021

Image tampering detection and localization have become an active area of research in the field of digital image forensics in recent times. This is due to the widespread of malicious image tampering. This study presents a new method for image tampering detection and localization that combines the advantages of dilated convolution, residual network, and UNET Architecture. Using the UNET architecture as a backbone, we built the proposed network from two kinds of residual units, one for the encoder path and the other for the decoder path. The residual units help to speed up the training process and facilitate information propagation between the lower layers and the higher layers which are often difficult to train. To capture global image tampering artifacts and reduce the computational burden of the proposed method, we enlarge the receptive field size of the convolutional kernels by adopting dilated convolutions in the residual units used in building the proposed network. In contrast to existing deep learning methods, having a large number of layers, many network parameters, and often difficult to train, the proposed method can achieve excellent performance with a fewer number of parameters and less computational cost. To test the performance of the proposed method, we evaluate its performance in the context of four benchmark image forensics datasets. Experimental results show that the proposed method outperforms existing methods and could be potentially used to enhance image tampering detection and localization.

• International Journal of Computer Science & Network Security
• /
• v.21 no.8
• /
• pp.238-246
• /
• 2021

The rapid rise of the Internet and social media has resulted in a large number of text-based reviews being placed on sites such as social media. In the age of social media, utilizing machine learning technologies to analyze the emotional context of comments aids in the understanding of QoS for any product or service. The classification and analysis of user reviews aids in the improvement of QoS. (Quality of Services). Machine Learning algorithms have evolved into a powerful tool for analyzing user sentiment. Unlike traditional categorization models, which are based on a set of rules. In sentiment categorization, Bidirectional Long Short-Term Memory (BiLSTM) has shown significant results, and Convolution Neural Network (CNN) has shown promising results. Using convolutions and pooling layers, CNN can successfully extract local information. BiLSTM uses dual LSTM orientations to increase the amount of background knowledge available to deep learning models. The suggested hybrid model combines the benefits of these two deep learning-based algorithms. The data source for analysis and classification was user reviews of Indian Railway Services on Twitter. The suggested hybrid model uses the Keras Embedding technique as an input source. The suggested model takes in data and generates lower-dimensional characteristics that result in a categorization result. The suggested hybrid model's performance was compared using Keras and Word2Vec, and the proposed model showed a significant improvement in response with an accuracy of 95.19 percent.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.5
• /
• pp.1778-1797
• /
• 2021

Document image binarization is an important pre-processing step in document analysis and archiving. The state-of-the-art models for document image binarization are variants of encoder-decoder architectures, such as FCN (fully convolutional network) and U-Net. Despite their success, they still suffer from three limitations: (1) reduced feature map resolution due to consecutive strided pooling or convolutions, (2) multiple scales of target objects, and (3) reduced localization accuracy due to the built-in invariance of deep convolutional neural networks (DCNNs). To overcome these three challenges, we propose an improved semantic segmentation model, referred to as DP-LinkNet, which adopts the D-LinkNet architecture as its backbone, with the proposed hybrid dilated convolution (HDC) and spatial pyramid pooling (SPP) modules between the encoder and the decoder. Extensive experiments are conducted on recent document image binarization competition (DIBCO) and handwritten document image binarization competition (H-DIBCO) benchmark datasets. Results show that our proposed DP-LinkNet outperforms other state-of-the-art techniques by a large margin. Our implementation and the pre-trained models are available at https://github.com/beargolden/DP-LinkNet.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 1998.06b
• /
• pp.189-194
• /
• 1998

We introduce a new zerotree scheme that effectively exploits the inter-scale self-similarities found in the octave decomposition by a wavelet transform. A zerotree is useful to efficiently code wavelet coefficients and its efficiency was proved by Shapiro's EZW. In the coding scheme, wavelet coefficients are symbolized and entropy-coded for more compression. The entropy per symbol is determined from the produced symbols and the final coded size is calculated by multiplying the entropy and the total number of symbols. In this paper, were analyze produced symbols from the EZW and discuss the entropy per symbol. Since the entropy depends on the produced symbols, we modify the procedure of symbolic streaming out for the purpose. First, we extend the relation between a parent and children used in the EZW to raise a probability that a significant parent has significant children. The proposed relation is flexibly extended according to the fact that a significant coefficient is highly addressed to have significant coefficients in its neighborhood. The extension way is reasonable because an image is decomposed by convolutions with a wavelet filter and thus neighboring coefficients are not independent with each other.