• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.4B
• /
• pp.202-209
• /
• 2005

In this paper, a dynamic TCP-friendly rate control (TFRC) is proposed to adjust the coding rates according to the channel characteristics of the wireless-to-wired network consisting of wireless first-hop channel. To avoid the throughput degradation of multimedia flows traveling through wireless lint the proposed rate control system employs a new wireless loss differentiation algorithm (LDA) using packet loss statistics. This method can produce the TCP-friendly rates while sharing the backbone bandwidth with TCP flows over the wireless-to-wired network. Experimental results show that the proposed rate control system can eliminate the effect of wireless losses in flow control of TFRC and substantially reduce the abrupt quality degradation of the video streaming caused by the unreliable wireless link status.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.4
• /
• pp.881-886
• /
• 2018

Over the last several decades, the advances in computer engineering technology have led a new paradigm of data-intensive research in the field of scientific applications. A collaborative work environment for storing, sharing and analyzing data is required for researchers of geographical distance. The Korean government supports the Research & Education network(KREONET) and the Global Science experimental Data hub Center(GSDC) to strengthen the nation's competitiveness. The KREONET has upgraded its backbone to 100Gbps to accommodate demand to transfer data fast among the global major experimental sites. This paper introduces the test result between high performance nodes reserved for big data transfer.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.958-961
• /
• 2019

Convolutional neural network(CNN)의 등장으로 얼굴 표현 인식 연구는 많은 발전을 이루었다. 그러나, 기존의 CNN 접근법은 미리 학습된 훈련모델에서 Multiple-level 의 의미적 맥락을 포함하지 않는 Attention-embedded 문제가 발생한다. 사람의 얼굴 감정은 다양한 근육의 움직임과 결합에 기초하여 관찰되며, CNN 에서 딥 레이어의 산출물로 나온 특징들의 결합은 많은 서브샘플링 단계를 통해서 class 구별와 같은 의미 정보의 손실이 일어나기 때문에 전이 학습을 통한 올바른 훈련 모델 생성이 어렵다는 단점이 있다. 따라서, 본 논문은 Backbone 네트워크의 Multi-level 특성에서 Channel-wise Attention 통합 및 의미 정보를 포함하여 높은 인식 성능을 달성하는 Densely Backwarnd Attention(DBA) CNN 방법을 제안한다. 제안하는 기법은 High-level 기능에서 채널 간 시멘틱 정보를 활용하여 세분화된 시멘틱 정보를 Low-level 버전에서 다시 재조정한다. 그런 다음, 중요한 얼굴 표정의 묘사를 분명하게 포함시키기 위해서 multi-level 데이터를 통합하는 단계를 추가로 실행한다. 실험을 통해, 제안된 접근방법이 정확도 79.37%를 달성 하여 제안 기술이 효율성이 있음을 증명하였다.

• 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.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.18 no.4
• /
• pp.1-11
• /
• 2022

Recently, research on license plate recognition, which is a core technology of an intelligent transportation system(ITS), is being actively conducted. In this paper, we propose a method to extract numbers and characters from low-quality license plate images by applying the YOLOv4 algorithm. YOLOv4 is a one-stage object detection method using convolution neural network including BACKBONE, NECK, and HEAD parts. It is a method of detecting objects in real time rather than the previous two-stage object detection method such as the faster R-CNN. In this paper, we studied a method to directly extract number and character regions from low-quality license plate images without additional edge detection and image segmentation processes. In order to evaluate the performance of the proposed method we experimented with 500 license plate images. In this experiment, 350 images were used for training and the remaining 150 images were used for the testing process. Computer simulations show that the mean average precision of detecting number and character regions on vehicle license plates was about 93.8%.

• Journal of the Korea Society for Simulation
• /
• v.19 no.4
• /
• pp.89-98
• /
• 2010

In the heterogeneous system of various wired or wireless network with IP-based backbone, the continuities of agreedon QoS for multimedia services should be guaranteed regardless of network types and terminal mobility through seamless vertical handover. This paper proposes a QoS provisioning mechanism called D-ISHO which guarantees the continuities of agreed-on QoS and seamless for multimedia services by considering both such characteristics as delay, loss rate and jitter per each service and such status as available band-width, call arrival rate and data transmission rate during the vertical handover. Simulation is done for performance analysis with the measure of handover failure rate and packet loss rate.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.625-627
• /
• 2022

In the construction and development of modern industrial production technology, the traditional technology management mode is faced with many problems such as low qualification rates and high application costs. In the research, an improved workpiece defect detection method based on deep learning is proposed, which can control the application cost and improve the detection efficiency of irregular defects. Based on the research of the current situation of deep learning applications, this paper uses the improved Faster R-CNN network structure model as the core detection algorithm to automatically locate and classify the defect areas of the workpiece. Firstly, the robustness of the model was improved by appropriately changing the depth and the number of channels of the backbone network, and the hyperparameters of the improved model were adjusted. Then the deformable convolution is added to improve the detection ability of irregular defects. The final experimental results show that this method's average detection accuracy (mAP) is 4.5% higher than that of other methods. The model with anchor size and aspect ratio (65,129,257,519) and (0.2,0.5,1,1) has the highest defect recognition rate, and the detection accuracy reaches 93.88%.

• Proceedings of the IEEK Conference
• /
• 2004.06a
• /
• pp.229-232
• /
• 2004

With the internet(World Wide Web) widespread among people after 1995, Information Communication Network, which is basic structure of the internet, is rapidly changing into fiber optics communication. Voice-centered Circuit Traffic is rapidly changing into Internet-centered Packet Traffic and Optics Communication Technology can transmit 1Tb/s thousands kilometers away. The information conveyed through the internet is being diversified - from simple information based on text to bulky multimedia information such as auido & video data. So the demand for higher speed, high storage Router is on the crease. With the continued increase of Internet Traffic, Router for Core Backbone System will be much more needed.

• Journal of the Korean Institute of Telematics and Electronics A
• /
• v.28A no.12
• /
• pp.1-11
• /
• 1991

Multi rate circuit switching ISDN which assumes a star topology consisting of LANs(local area networks) interconnected through a backbone MAN (metropolitan area network) is considered. The stations are classified into two categories data and voice stations they are characterized by different arrival rate and session-length (holding time) statistics. We first model the data and voice terminals as finite-source stations and derive analytically exact expressions for the end-to-end blocking probabilities. For more exact evaluation, these probabilities are calculated in terms of three components:circuit blocking probability, destination busy probability and concurrent busy probability. For large size systems, we develop an develop an iterative algorithm that provides a computationally efficient and fast method for the calculation of the end-to-end blocking probabilities.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 2004.04a
• /
• pp.219-224
• /
• 2004

In recent years, as the demands of VBNS(Very high speed Backbone Network Service) and VDSL(Very high-data rate Digital Subscriber Line) increase, the development of kernel parts of optical communication such as PLC(Planar Light Circuit), Coupler, WDM elements increase. The alignment and the attachment technology are very important to fabricate the optical elements for communication. In this paper, the mechanism of rotational stage, the contact sensing unit, and integrated control circuit for the optical alignment system are studied.