• Annual Conference of KIPS
• /
• 2022.11a
• /
• pp.21-23
• /
• 2022

ARM 아키텍처를 사용하는 임베디드 시스템에서 int8, fp16, fp32 데이터를 조합하여 c/c++로 작성된 mixed precision CNN 을 실행시키기 위한 프레임워크 구성으로, 네트워크의 레이어마다 다른 정밀도를 사용하여 네트워크 경량화 및 추론 정확도 향상을 위한 최적의 설정을 탐색하는 실험 및 분석이 가능토록 하는 것을 목적으로 한다. 주요 구성은 network forwarding 중 레이어의 입력이 레이어에 설정된 정밀도와 다를 경우 실행되는 양자화/반양자화를 c/c++로 바인딩된 tensorflow 의 quantization 모듈을 사용하여 진행하고 ARM 시스템에서 c/c++의 fp16 을 사용하기 위해 fp16 를 컴파일이 가능한 ARM compiler 를 사용하는 프레임워크를 제안한다.

• Information Systems Review
• /
• v.4 no.2
• /
• pp.41-58
• /
• 2002

This paper is to show the meaningfulness of applying Information Technology Architecture (ITA) to the modeling of a nation-wide e-Business architecture framework. After reviewing various architecture frameworks we suggested a unique architecture framework called VMT (Views, Models, and Time-frames). VMT represents five views of e-Business stakeholders; six models of data, function, network, agent, event, and rule; and three timeframes of short-term, mid-term, and long-term. VMT is an extension and unification of popular frameworks including Zachman's framework that has international recognition and use, C4ISR architecture framework of US DoD's, and Federal Enterprise Architecture Framework (FEAF) of the US Federal Government's.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2009.06a
• /
• pp.745-746
• /
• 2009

• Journal of Korea Game Society
• /
• v.15 no.4
• /
• pp.199-206
• /
• 2015

This paper introduces a serious game called freCman developed for C programming language beginners. Since key words, syntax, and programming structure of C programming language are unfamiliar for them, they feel uneasy and have many difficulties to study. We developed three games such as shooting star C, finding hidden errors, unscrambling C codes through which C beginners can study C language easily. Also we developed CTS(Code to Speech) which speeches C source codes like English statements so that C beginners can be familiar with C key words and statements. To prove effectiveness of the freCman, some experiments have been conducted with C language beginners. Experiment results show that the freCman helps beginners studying C programming language much.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.35 no.8
• /
• pp.951-956
• /
• 2011

Inconel alloys are generally used as steam generator tubes in nuclear power plants. These alloys are highnickel chromium alloys that exhibit excellent resistance to aqueous corrosion. In this paper, the effects of elevated temperatures such as an operating temperature of $320^{\circ}C$ on the fretting fatigue behavior of inconel 600 and 690. We observed that the plain and fretting fatigue limits at $320^{\circ}C$ were slightly lower than those at room temperature. The frictional forces varied depending on the number of load cycles. After each test, we studied the fretting fatigue mechanisms via SEM observations. These results can be used for structural integrity evaluations at elevated temperatures and for studying fretting damage in steam generator systems.

• The KIPS Transactions:PartC
• /
• v.15C no.6
• /
• pp.547-556
• /
• 2008

Recently, with the advance of computing and networking technique, the high speed internet becomes widespread, however, it is still hard job to do streaming the media which requires high network bandwidth over the internet. Previous researches for streaming over the internet mainly proposed techniques that controls the QoS(Quality of Service) of the media in proportion to the network status. Though, this could be the solution for the service provider while the service user who wants constant QoS may not satisfy with variable QoS. In the paper, we propose a network adaptive prefetching technique, NAP, for guarantee of constant QoS. The NAP prefetches frames by increasing the frame transmission rate while the available network bandwidth is high. The NAP uses the prefetched frames to guarantee the QoS while the available network bandwidth is low. The experiment result shows that the proposed NAP could guarantee the constant QoS by prefetching the frames adaptively to the network bandwidth with the characteristic of video stream.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.11C
• /
• pp.1066-1072
• /
• 2007

In this paper, we propose a fast multiple reference frame selection method for motion estimation and compensation in video coding. Reference frames selected as an optimal reference frame by variable block sizes motion estimation have the statistical characteristic that was based on block size. Using the statistical characteristic, reference frames for smaller block size motion estimation can be selected from reference frame which was decided as an optimal one for the upper layer block size. Simulation results show that the proposal method decreased the computations about 60%. Nevertheless, PSNR and bit rate were almost same as the performances of original H.264 multiple reference motion estimation.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.35C no.11
• /
• pp.48-56
• /
• 1998

In this paper, a moving object detection and face region extraction algorithm which can be used in video monitoring systems is presented. The proposed algorithm is composed of two stages. In the first stage, each frame of an input video sequence is analyzed using three measures which are based on image pixel difference. If the current frame contains moving objects, their skin regions are extracted using color and frame difference information in the second stage. Since the proposed algorithm does not rely on computationally expensive features like optical flow, it is well suited for real-time applications. Experimental results tested on various sequences have shown the robustness of the proposed algorithm.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.8C
• /
• pp.764-771
• /
• 2009

In this paper, we propose a fast multiple reference frame selection method using neighboring motion vectors and the correlation of inter-mode blocks in H.264 video coding. Using the property that an optimal reference frame has the statistical characteristics, the motion estimation of small Inter-block size is estimated from the motion estimation of the larger Inter-block size. Simulation results show that the proposal method decreased the computations about 91%. Without the sacrifice of coding performance, comparing to the H.264 multiple reference picture coding.

• Journal of Advanced Navigation Technology
• /
• v.14 no.4
• /
• pp.553-561
• /
• 2010

This paper proposed the face tracking method which can be effectively applied to the robot's vision system. The proposed algorithm tracks the facial areas after detecting the area of video motion. Movement detection of video images is done by using median filter and erosion and dilation operation as a method for removing noise, after getting the different images using two continual frames. To extract the skin color from the moving area, the color information of sample images is used. The skin color region and the background area are separated by evaluating the similarity by generating membership functions by using MIN-MAX values as fuzzy data. For the face candidate region, the eyes are detected from C channel of color space CMY, and the mouth from Q channel of color space YIQ. The face region is tracked seeking the features of the eyes and the mouth detected from knowledge-base. Experiment includes 1,500 frames of the video images from 10 subjects, 150 frames per subject. The result shows 95.7% of detection rate (the motion areas of 1,435 frames are detected) and 97.6% of good face tracking result (1,401 faces are tracked).