• Journal of Broadcast Engineering
• /
• v.26 no.6
• /
• pp.748-757
• /
• 2021

In this paper, we propose a gaze estimation network in which eye landmark position detection and gaze direction vector estimation are integrated into one deep learning network. The proposed network uses the Stacked Hourglass Network as a backbone structure and is largely composed of three parts: a landmark detector, a feature map extractor, and a gaze direction estimator. The landmark detector estimates the coordinates of 50 eye landmarks, and the feature map extractor generates a feature map of the eye image for estimating the gaze direction. And the gaze direction estimator estimates the final gaze direction vector by combining each output result. The proposed network was trained using virtual synthetic eye images and landmark coordinate data generated through the UnityEyes dataset, and the MPIIGaze dataset consisting of real human eye images was used for performance evaluation. Through the experiment, the gaze estimation error showed a performance of 3.9, and the estimation speed of the network was 42 FPS (Frames per second).

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.12B
• /
• pp.1722-1727
• /
• 2011

The Korea Army are developing the TICN(Tactical Information Communication Network) which is the next army tactical communication system. The TICN tactical backbone's ability is very important. So analysis of the standard protocol for the data transmission guarantee are obliged. For the TICN tactical backbone, previous paper analysed routing protocol efficiency index and compared to superiority of the OSPF and IS-IS protocols in the tactical environment and conclusively IS-IS protocol was more suitable for TICN backbone routing protocol. And We got the result that It could improve the performance of IS-IS protocol as a TICN tactical backbone protocol under the tactical environments by adjusting some time parameters of IS-IS protocol. This paper modified and designed the IS-IS protocol in order to support the QoS strategies of TICN and analyzed it's adaptability through simulation.

• Proceedings of the IEEK Conference
• /
• 2003.07a
• /
• pp.73-76
• /
• 2003

The capacity of backbone networks has largely kept pace with the tremendous of growth of Internet traffic. But there has been little progress in the access networks. In this paper, we introduce the access network that is able to offers high bandwidth and its simple protocol. The network architecture is basically multiple ring and use WDM(Wavelength Division Multiplexing), SCM (SubCarrier Multiplexing), and static WADM (Wavelenth Add/Drop Multiplexer) for simplicity, low cost, and offering high bandwidth.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10e
• /
• pp.298-300
• /
• 2002

급격한 기술의 발전 및 신기술의 등장에 따라 국가적 차원에서 고속, 대량의 데이터를 처리하는 네트워크를 구축할 필요성이 발생하였다. 이에 고속 통신망을 구축, 운영 중에 있으나, 현재의 network은 망의 안정성, 생존성 확보를 위하여 다수의 장거리 전용회선을 사용하고 있다. 본 논문에서는 현재의 network 구조에 피터슨 그래프를 이용하여 약간의 수정을 가하여 기존 운영중인 망에서 생존성을 보장하고 경제성을 향상시키는 효율적 망 활용 방법을 제시한다.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.47 no.7
• /
• pp.135-142
• /
• 2010

Wireless mesh network, which is an access network technology, adopts ubiquitous features of ad-hoc network that includes capabilities of self-configuration and self-management. This paper proposes a scheme which enables nodes along route in wireless mesh network to authenticate data and verify data integrity. The scheme distinguishes infra-node, which is a network device used to form mesh network, and user node in ad-hoc network, which operates functions as a sender, receiver or relayer, to deploy different authentication scheme. That is, hop-based authentication scheme along route forming wireless backbone differs from authentication scheme for user nodes in route over MANET. The proposed scheme is less complex than previously proposed schemes from the repects of security setup procedures and managements. In addition, the scheme is able to reduce transmission delay from a source to a destination owing to fast authentication over wireless backbone.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.12
• /
• pp.4439-4455
• /
• 2021

Compared with two-stage object detection algorithms, one-stage algorithms provide a better trade-off between real-time performance and accuracy. However, these methods treat the intermediate features equally, which lacks the flexibility to emphasize meaningful information for classification and location. Besides, they ignore the interaction of contextual information from different scales, which is important for medium and small objects detection. To tackle these problems, we propose an image pyramid network based on dual attention mechanism (DAIPNet), which builds an image pyramid to enrich the spatial information while emphasizing multi-scale informative features based on dual attention mechanisms for one-stage object detection. Our framework utilizes a pre-trained backbone as standard detection network, where the designed image pyramid network (IPN) is used as auxiliary network to provide complementary information. Here, the dual attention mechanism is composed of the adaptive feature fusion module (AFFM) and the progressive attention fusion module (PAFM). AFFM is designed to automatically pay attention to the feature maps with different importance from the backbone and auxiliary network, while PAFM is utilized to adaptively learn the channel attentive information in the context transfer process. Furthermore, in the IPN, we build an image pyramid to extract scale-wise features from downsampled images of different scales, where the features are further fused at different states to enrich scale-wise information and learn more comprehensive feature representations. Experimental results are shown on MS COCO dataset. Our proposed detector with a 300 × 300 input achieves superior performance of 32.6% mAP on the MS COCO test-dev compared with state-of-the-art methods.

• Journal of the Semiconductor & Display Technology
• /
• v.22 no.1
• /
• pp.11-17
• /
• 2023

This paper proposes a lightweight deep learning network for detecting an image splicing forgery. The research on image forgery detection using CNN, a deep learning network, and research on detecting and localizing forgery in pixel units are in progress. Among them, CAT-Net, which learns the discrete cosine transform coefficients of images together with images, was released in 2022. The DCT coefficients presented by CAT-Net are combined with the JPEG artifact learning module and the backbone model as pre-learning, and the weights are fixed. The dataset used for pre-training is not included in the public dataset, and the backbone model has a relatively large number of network parameters, which causes overfitting in a small dataset, hindering generalization performance. In this paper, this learning module is designed to learn the characterization depending on the DCT domain in real-time during network training without pre-training. The DCT RRU-Net proposed in this paper is a network that combines RRU-Net which detects forgery by learning only images and JPEG artifact learning module. It is confirmed that the network parameters are less than those of CAT-Net, the detection performance of forgery is better than that of RRU-Net, and the generalization performance for various datasets improves through the network architecture and training method of DCT RRU-Net.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.4
• /
• pp.618-636
• /
• 1997

DQDB(Distributed Queue Dual Bus) protocol, the IEEE 802.6 standard protocol for metropolitan area networks, does not fully take advantage of the capabilities of dual bus architecture. Although fairness in bandwidth distribution among nodes is improved when using so called the bandwidth balancing mechanism, the protocol requires a considerable amount of time to adjust to changes in the network load. Additionally, the bandwidth balancing mechanism leaves a portion of the available bandwidth unused. In a high-speed backbone network, each node may act as a bridge/ router which connects several LANs as well as hosts. However, Because the existence of high speed LANs becomes commonplace, the congestionmay occur on a node because of the limitation on access rate to the backbone network and on available buffer spaces. to release the congestion, it is desirable to install some congestion control algorithm in the node. In this paper, we propose an efficient congestion control mechanism and fair and waster-free MAC protocol for dual bus network. In this protocol, all the buffers in the network can be shared in such a way that the transmission rate of each node can be set proportional to its load. In other words, a heavily loaded node obtains a larger bandwidth to send the sements so tht the congestion can be avoided while the uncongested nodes slow down their transmission rate and store the incoming segments into thier buffers. this implies that the buffers on the network can be shared dynamically. Simulation results show that the proposed probotol significantly reduces the segment queueing delay of a heavily loaded node and segment loss rate when compared with original DQDB. And it enables an attractive high throughput in the backbone network. Because in the proposed protocol, each node does not send a requet by the segment but send a request one time in the meaning of having segments, the frequency of sending requests is very low in the proposed protocol. so the proposed protocol signigificantly reduces the segment queuing dely. and In the proposed protocol, each node uses bandwidth in proportion to its load. so In case of limitation of available buffer spaces, the proposed protocol reduces segment loss rate of a heavily loaded node. Bandwidth balancing DQDB requires the wastage of bandwidth to be fair bandwidth allocation. But the proposed DQDB MAC protocol enables fair bandwidth without wasting bandwidth by using bandwidth one after another among active nodes.

• Journal of IKEEE
• /
• v.27 no.3
• /
• pp.345-349
• /
• 2023

In this paper, we propose a study on the development of deep learning structure for defective pixel detection of next-generation smart LED display board using imaging device. In this research, a technique utilizing imaging devices and deep learning is introduced to automatically detect defects in outdoor LED billboards. Through this approach, the effective management of LED billboards and the resolution of various errors and issues are aimed. The research process consists of three stages. Firstly, the planarized image data of the billboard is processed through calibration to completely remove the background and undergo necessary preprocessing to generate a training dataset. Secondly, the generated dataset is employed to train an object recognition network. This network is composed of a Backbone and a Head. The Backbone employs CSP-Darknet to extract feature maps, while the Head utilizes extracted feature maps as the basis for object detection. Throughout this process, the network is adjusted to align the Confidence score and Intersection over Union (IoU) error, sustaining continuous learning. In the third stage, the created model is employed to automatically detect defective pixels on actual outdoor LED billboards. The proposed method, applied in this paper, yielded results from accredited measurement experiments that achieved 100% detection of defective pixels on real LED billboards. This confirms the improved efficiency in managing and maintaining LED billboards. Such research findings are anticipated to bring about a revolutionary advancement in the management of LED billboards.

• Journal of KIISE:Information Networking
• /
• v.27 no.2
• /
• pp.219-225
• /
• 2000

In order to provide a proper mobile internet service, Mobile IP is necessary to support IP mobility. Service network should be a backbone network among mobile agents which support mobility, and MPLS(Multi-protocol Label Switching) of IETF(Internet Engineering Task Force) is being considered as a backbone network because of its speed, scalability and the excellent service capability. MPLS, however, doesn't provide a way to support the mobility of the nodes. In this paper, we present an efficient IMT-2000 network architecture using MPLS to handle Mobile IP. The proposed architecture combines the MPLS label distribution and Mobile IP registration. It doesn't use the layer 3 encapsulation, instead it uses layer 2 for tunneling the data, reduces the size of the header, and it can tunnel the data without delay, which is needed to look up the mobility binding list, as a result.