• KNOM Review
• /
• v.22 no.1
• /
• pp.20-29
• /
• 2019

Recently, various machine learning based traffic classification methods are focused on detecting malicious network traffic. In this paper, convolutional neural network based malicious network traffic classification method is introduced and its performance is evaluated. In order to utilize the convolutional neural network which is excellent in analyzing images, a image transform method from important information of network traffic to a standardized image is proposed, and the transformed images are used as learning input of a CNN network traffic classifier. By using the real network traffic dataset, the proposed image transform method and CNN based network traffic classification method are evaluated. Especially, under various configurations of CNN, the performance of the proposed method is evaluated.

• KIISE Transactions on Computing Practices
• /
• v.24 no.2
• /
• pp.93-98
• /
• 2018

Recommendation systems are used to provide items of interests for users to maximize a company's profit. Matrix factorization is frequently used by recommendation systems, based on an incomplete user-item rating matrix. However, as the number of items and users increase, it becomes difficult to make accurate recommendations due to the sparsity of data. To overcome this drawback, the use of text data related to items was recently suggested for matrix factorization algorithms. Furthermore, a word-level convolutional neural network was shown to be effective in the process of extracting the word-level features from the text data among these kinds of matrix factorization algorithms. However, it involves a large number of parameters to learn in the word-level convolutional neural network. Thus, we propose a matrix factorization algorithm which utilizes a character-level convolutional neural network with which to extract the character-level features from the text data. We also conducted a performance study with real-life datasets to show the effectiveness of the proposed matrix factorization algorithm.

• The Korean Journal of Applied Statistics
• /
• v.32 no.4
• /
• pp.573-591
• /
• 2019

Classification criteria for Korean alphabet (Hangul) fonts are undeveloped in comparison to numerical classification systems for Roman alphabet fonts. This study finds important features that distinguish typeface styles in order to help develop numerical criteria for Hangul font classification. We find features that determine the characteristics of the two different styles using a convolutional neural network to create a model that analyzes the learned filters as well as distinguishes between serif and sans-serif styles.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.06a
• /
• pp.205-206
• /
• 2018

본 논문에서는 디지털 홀로그램 영상을 Fresnelet 변환을 하여 상관도를 확인할 수 있는 데이터로 바꾸고, 컨볼루션 오토인코더(Convolutional Autoencoder, CAE)를 이용해 압축하고 생성하는 방법을 제안한다. 컨볼루션 계층과 채널 수가 다른 2개의 네트워크로 실험한다. CAE의 인코더를 수행해 영상을 압축하고 디코더를 통해 복원한다. 원본 영상의 Fresnelet 영역과 2개의 네트워크를 진행하여 생성된 Fresnelet 영역을 다시 역 Fresnelet하여 압축률에 따른 PSNR을 비교, 분석한다.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.6
• /
• pp.861-868
• /
• 2023

One of the key technologies in providing data analysis in the deep learning while maintaining security is fully homomorphic encryption. Due to constraints in operations on fully homomorphically encrypted data, non-arithmetic functions used in deep learning must be approximated by polynomials. Until now, the degrees of approximation polynomials with composite minimax polynomials have been uniformly set across layers, which poses challenges for effective network designs on fully homomorphic encryption. This study theoretically proves that setting different degrees of approximation polynomials constructed by composite minimax polynomial in each layer does not pose any issues in the inference on convolutional neural networks.

• 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.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.641-643
• /
• 2023

Deep Learning(DL)을 사용한 호흡음의 자동 분석은 폐 질환의 조기 진단에 중추적인 역할을 한다. 그러나 현재의 DL 방법은 종종 호흡음의 공간적 및 시간적 특성을 분리하여 검사하기 때문에 한계가 있다. 본 연구는 컨볼루션 연산을 통해 공간적 특징을 캡처하고 시간 컨볼루션 네트워크를 사용하여 이러한 특징의 공간적-시간적 상관 관계를 활용하는 새로운 DL 프레임워크를 제한한다. 제안된 프레임워크는 앙상블 학습 접근법 내에 컨볼루션 네트워크를 통합하여 폐음 녹음에서 호흡 이상 및 질병을 검출하는 정확도를 크게 향상시킨다. 잘 알려진 ICBHI 2017 챌린지 데이터 세트에 대한 실험은 제안된 프레임워크가 호흡 이상 및 질병 검출을 위한 4-Class 작업에서 비교모델 성능보다 우수함을 보여준다. 특히 민감도와 특이도를 나타내는 점수 메트릭 측면에서 최대 45.91%와 14.1%의 개선이 이진 및 다중 클래스 호흡 이상 감지 작업에서 각각 보여준다. 이러한 결과는 기존 기술보다 우리 방법의 두드러진 이점을 강조하여 호흡기 의료 기술의 미래 혁신을 주도할 수 있는 잠재력을 보여준다.

• Journal of Broadcast Engineering
• /
• v.23 no.3
• /
• pp.369-382
• /
• 2018

We propose a single-shot high dynamic range (HDR) imaging algorithm using a deep convolutional neural network (CNN) for row-wise varying exposures in a single image. The proposed algorithm restores missing information resulting from under- and/or over-exposed pixels in an input image and reconstructs the raw radiance map. The main contribution of this work is the development of a loss function for the CNN employing the human visual system (HVS) properties. Then, the HDR image is obtained by applying a demosaicing algorithm. Experimental results demonstrate that the proposed algorithm provides higher-quality HDR images than conventional algorithms.

• Smart Media Journal
• /
• v.7 no.3
• /
• pp.16-21
• /
• 2018

This paper presents the design and implementation of a deep learning-based automated weed detector on onion fields. The system is based on a Convolutional Neural Network that specifically selects proposed regions. The detector initiates training with a dataset taken from agricultural onion fields, after which candidate regions with very high probability of suspicion are considered weeds. Non-maximum suppression helps preserving the less overlapped bounding boxes. The dataset collected from different onion farms is evaluated with the proposed classifier. Classification accuracy is about 99% for the dataset, indicating the proposed method's superior performance with regard to weed detection on the onion fields.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.5
• /
• pp.801-808
• /
• 2022

Research on how to embed knowledge in large-scale Linked Data and apply neural network models for entity matching is relatively scarce. The most fundamental problem with this is that different labels lead to lexical heterogeneity. In this paper, we propose an extended GCN (Graph Convolutional Network) model that combines re-align structure to solve this lexical heterogeneity problem. The proposed model improved the performance by 53% and 40%, respectively, compared to the existing embedded-based MTransE and BootEA models, and improved the performance by 5.1% compared to the GCN-based RDGCN model.