• 제목/요약/키워드: EfficientNetB2

검색결과 34건 처리시간 0.027초

DEA를 이용한 B2B 온라인 쇼핑몰 상품관리 효율성 증대 방안 (Improving the Utilization and Efficiency of B2B Online Store using DEA)

  • 구승환;박현기;장성용
    • 한국산학기술학회논문지
    • /
    • 제15권7호
    • /
    • pp.4237-4245
    • /
    • 2014
  • 본 연구는 DEA 모형을 이용하여 B2B 온라인 쇼핑몰의 상품을 효율적으로 분류하고 운영하는 방법을 제시하였다. 이를 위해 실제 M사의 자료를 활용하여 분석을 수행하였으며, 카테고리를 분류하여 업무부하량을 산정하였다. 연구결과, DEA를 활용하여 상품을 분리해 관리할 경우 기존에 비해 업무 부하량이 평준화되는 것을 알 수 있었다. 또한 A분류에는 순이익이 가장 높은 카테고리들로 구성되어 있어 회사에서 주력으로 관리해야할 분류임을 알 수 있었다. B와 C분류의 업무는 상대적으로 중요도가 낮은 것으로 구성되어 있으며, 기존보다 업무 부하량도 줄어든 것을 알 수 있었다. 따라서 본 연구결과를 실제 업무에 반영한다면 효율적인 운영이 가능 할 것으로 사료된다.

6LoWPAN 노드와 외부 노드의 통신 시에 오버헤드 감소 방법 (Overhead Reduction Methods in Communication between 6LoWPAN and External Node)

  • 최대인;;박종탁;강현국
    • 한국통신학회논문지
    • /
    • 제36권5B호
    • /
    • pp.437-442
    • /
    • 2011
  • 6LoWPAN은 IETF Working Group으로 IEEE 802.15.4 상에서 IPv6 패킷을 전송하는 기술의 표준화를 진행하고 있다. 6LoWPAN WG은 현재 2개의 RFC 문서를 제정하였다. 이 중에 RFC 4944는 단편화, 재조립, 헤더 압축의 기술이 포함되어 있다. 본 논문은 6LoWPAN 노드가 외부 노드와 통신 할 경우에 효율적인 통신 메카니즘을 제안 한다. 이 통신 메카니즘은 6LoWPAN과 외부 네트워크 사이의 게이트웨이가 각 노드 사이에서 프락시 게이트웨이처럼 동작하는 방법을 사용 하였다. 성능 비교 및 분석을 위하여 QualNet으로 시뮬레이션을 하였다. 이를 통하여 제안한 메카니즘과 기존의 RFC 4944의 방법을 비교하여 성능이 향상되었음을 확인하였다.

대학병원 영양부서 운영체계 변경의 비용.편익분석 (Cost , Benefit Analysis of Operation System Change in the Hospital Foodservice)

  • 김형미;양일선;박은철;임현숙
    • 대한영양사협회학술지
    • /
    • 제6권1호
    • /
    • pp.33-43
    • /
    • 2000
  • Environmental pressures from such sources an economic condition, the government and inter-institutional competition create managerial challenges. Economic pressures may be forcing dietetic dept, in hospital to utilize cost∙benefit analysis to assist them in their problem solving. Cost∙benefit analysis have been widely used in business, industry and many other fields with only limited application to foodservice. Due to the lack or this information the purposes of this study were to identify use of cost∙benefit analysis in hospital foodservice system to evaluate the economic efficiency of alternatives, and to make recommendation for operation system change. Using the cost∙benefit method, cash flows are separated into cost and benefits. For an alternative to be selected, indicators, such as NPV, benefit-cost ratio (B/C ratio) with 5% discount rate per annum. The sensitivity analysis was also conducted with difference rate 3%, 7% respectively and reduced employee payroll change. The result of this study can be summarized as follows : 1. The total cost of investment for operation system change was 390,570 thousand won and the total benefit through operation system change was 865,808 thousand won. 2. Net present value(NPV) for 5 years was 475,239 thousand won and benefit-cost ratio was 2.22. 3. In sensitivity analysis with different discount rate 3%, 7%, benefit-cost ratio was 2.25, 2.18 respectively, with total reduced employee payroll change, benefit-cost ratio was 2.86. In conclusion, total benefits were exceeded total costs. Therefore, the project of operation system change in hospital foodservice was found to be economically efficient.

  • PDF

사회적 기업의 자료포락분석(DEA)을 통한 경영효율성 평가 (Management Efficiency Estimation of Social Enterprises with Data Envelopment Analysis)

  • 이상연;임성묵;채명신
    • 산업경영시스템학회지
    • /
    • 제40권2호
    • /
    • pp.121-128
    • /
    • 2017
  • This paper was to evaluate social enterprises' management efficiency with Data Envelope Analysis (DEA). The data was based on the 168 social enterprises' of annual performance reports published in 2015. The research focused on to measure both financial efficiency and social impact of the companies simultaneously. To apply DEA, the paper classified the enterprises into seven types based on types of socal impacts which each company provides before the estimation of the efficiency. The research results showed that group D, which employes disadvantaged people, provides social services and shares resources was the most efficient group and had higest net worths in Pure Technical Efficiency. In contrast, Group B, which only employs social advantage people and provides social service, was the least efficient one. The research suggests a practical and efficient framework in measuring social enterprises' management efficiency, including both the financial performance and social impacts simultaneously with their self-publishing reports. Because the Korea Social Enterprise Promotion Agency does not open business reports which social enterprises submit each year, there are basic limitations on researchers attempting to analyse with data from all social enterprises in Korea. Thus, this study dealt with only 10% of the social enterprises which self-published their performance report on the Korea Social Enterprise Promotion Agency's web site. Regardless of these limitations, this study suggested substantial methods to estimate management efficiency with the self-published reports. Because self-publishing is increasing each year, it will be the main source of information for researchers in examining and evaluating social enterprises' financial performance or social contribution. The research suggests a practical and efficient framework in measuring social enterprises' management efficiency, including both the financial performance and social impacts simultaneously with their self-publishing reports. The research results suggest not only list of efficient enterprises but also methods of improvement for less efficient enterprises.

CT 정도관리를 위한 인공지능 모델 적용에 관한 연구 (Study on the Application of Artificial Intelligence Model for CT Quality Control)

  • 황호성;김동현;김호철
    • 대한의용생체공학회:의공학회지
    • /
    • 제44권3호
    • /
    • pp.182-189
    • /
    • 2023
  • CT is a medical device that acquires medical images based on Attenuation coefficient of human organs related to X-rays. In addition, using this theory, it can acquire sagittal and coronal planes and 3D images of the human body. Then, CT is essential device for universal diagnostic test. But Exposure of CT scan is so high that it is regulated and managed with special medical equipment. As the special medical equipment, CT must implement quality control. In detail of quality control, Spatial resolution of existing phantom imaging tests, Contrast resolution and clinical image evaluation are qualitative tests. These tests are not objective, so the reliability of the CT undermine trust. Therefore, by applying an artificial intelligence classification model, we wanted to confirm the possibility of quantitative evaluation of the qualitative evaluation part of the phantom test. We used intelligence classification models (VGG19, DenseNet201, EfficientNet B2, inception_resnet_v2, ResNet50V2, and Xception). And the fine-tuning process used for learning was additionally performed. As a result, in all classification models, the accuracy of spatial resolution was 0.9562 or higher, the precision was 0.9535, the recall was 1, the loss value was 0.1774, and the learning time was from a maximum of 14 minutes to a minimum of 8 minutes and 10 seconds. Through the experimental results, it was concluded that the artificial intelligence model can be applied to CT implements quality control in spatial resolution and contrast resolution.

2차원 단면 데이터로부터 복합 $G^{1}$ 자유곡면 생성 (Composite $G^{1}$ surface construction from 2D cross-sections)

  • 박형준;나상욱;배채열
    • 한국경영과학회:학술대회논문집
    • /
    • 대한산업공학회/한국경영과학회 2004년도 춘계공동학술대회 논문집
    • /
    • pp.44-47
    • /
    • 2004
  • This paper proposes an approach for composite surface reconstruction from 2D serial cross-sections, where the number of contours varies from section to section. In a triangular surface-based approach taken in most reconstruction methods, a triangular $G^{1}$ surface is constructed by stitching triangular patches over a triangular net generated from the compiled contours. In the proposed approach, the resulting surface is a composite $G^{1}$ surface consisting of three kinds of surfaces: skinned, surface is first represented by a B-spline surface approximating the serial contours of the skinned region and then serial contours of the skinned region and then transformed into a mesh of rectangular Bezier patches. On branched and capped regions, triangular $G^{1}$ surfaces are constructed so that the connections between the triangular surfaces and their neighboring surfaces are $G^{1}$ continuous. Since each skinned region is represented by an approximated rectangular $G^{2}$ surface instead of an interpolated triangular $G^{1}$ surface, the proposed approach can provide more visually pleasing surfaces and realize more efficient data reduction than the triangular surface-based approach. Some experimental results demonstrate its usefulness and quality.

  • PDF

M/W 중계 시스템 망의 주파수 조정을 위한 보호비 계산에 대한 연구 (A Study on Calculation of Protection Ratio for Frequency Coordination in Microwave Relay System Networks)

  • 서경환
    • 한국전자파학회논문지
    • /
    • 제17권2호
    • /
    • pp.139-147
    • /
    • 2006
  • 본 논문에서는 마이크로파 중계 시스템 망의 주파수 조정에 적용할 수 있는 효율적인 보호비 산출 방법을 제안하고 결과를 제시한다. 또한 인접 채널 보호비를 계산하기 위해 송신 스펙트럼 마스크와 수신 필터 특성에 관련된 통합 필터 변별도(NFD)를 고찰하였다. 보호비는 변조 방식에 따른 신호-대-잡음 비, 잡음-대-간섭 비, 다중 간섭 허용, 다중 경로 또는 강우 감쇠의 페이드 마진, 통합 필터 변별도의 변수들로 구성된다. 주파수 6.7 GHz, 64-QAM, 거리 60 km, BER $10^{-6}$에서 계산된 페이드 마진 및 보호비는 각각 41.1 및 75.2 dB를 얻을 수 있었다. 채널 대역폭 40 MHz의 NFD는 첫 번째 인접 채널에서 28.9 dB가 되며, 이로부터 첫 번째 인접 채널의 보호비는 46.3 dB가 됨을 알 수 있었다. 또한 실제 중계망의 적용을 위해 채널 대역폭 20 및 40 MHz를 갖는 이종 시스템간의 NFD 및 보호비도 고찰하였다. 제안된 방법은 계산의 용이성과 체계적 확장, 그리고 밀리미터파 중계망의 주파수 조정에도 동일한 개념을 적용할 수 있는 장점을 갖는다.

광대역 음성 부호화기용 선 스펙트럼 주파수 계수 양자화기 설계 (Design of the LSF Parameter Quantizer for the Wideband Speech Codec)

  • 지상현;강상원;윤병식
    • 한국음향학회지
    • /
    • 제20권4호
    • /
    • pp.29-34
    • /
    • 2001
  • 본 논문에서는 고품질 음성 서비스를 가능하게 하는 광대역 음성 부호화기의 선 스펙트럼 주파수 (line spectral frequency: ISF) 계수 양자화기를 설계하였다. 광대역 음성 부호화기를 위한 효율적인 LSF 계수 양자화기를 설계하기 위하여, 인접 프레임간의 상관도를 이용하였으며, 각 해당 프레임의 ISF 계수에 대한 양자화를 인접 프레임간 상관도가 높은 프레임과 상관도가 낮은 프레임으로 나누어 독립적으로 수행하였다. 인접 프레임간 상관도가 높은 프레임의 LSF계수 양자화를 위하여 예측 피라미드형 벡터 양자화기 (predictive pyramid vector quantizer: PPVQ)를 사용하여 양자화하였고, 상관도가 낮은 프레임의 LSF 계수는 피라미드형 벡터 양자화기 (PVQ)를 사용하여 양자화 하였다. PPVQ에서 예측기로 1차 AR 예측기를 사용하였다. 광대역 음성 부호화기를 위해 본 논문에서 설계된 UF 계수양자화기를 평균스펙트럼 왜곡(spectral distortion: SD) 성능 관점에서 실험한 결과, LSF계수 양자화에 할당된 비트가 프레임당 40비트일 때, 평균 SD값이 1 dB 내외이고, 2 dB 이상 및 4 dB 이상 outlier가 각각 3.87%및 0.01%인 transparent한 성능을 얻을 수 있었다.

  • PDF

Performance Analysis of Sensor Systems for Space Situational Awareness

  • Choi, Eun-Jung;Cho, Sungki;Jo, Jung Hyun;Park, Jang-Hyun;Chung, Taejin;Park, Jaewoo;Jeon, Hocheol;Yun, Ami;Lee, Yonghui
    • Journal of Astronomy and Space Sciences
    • /
    • 제34권4호
    • /
    • pp.303-314
    • /
    • 2017
  • With increased human activity in space, the risk of re-entry and collision between space objects is constantly increasing. Hence, the need for space situational awareness (SSA) programs has been acknowledged by many experienced space agencies. Optical and radar sensors, which enable the surveillance and tracking of space objects, are the most important technical components of SSA systems. In particular, combinations of radar systems and optical sensor networks play an outstanding role in SSA programs. At present, Korea operates the optical wide field patrol network (OWL-Net), the only optical system for tracking space objects. However, due to their dependence on weather conditions and observation time, it is not reasonable to use optical systems alone for SSA initiatives, as they have limited operational availability. Therefore, the strategies for developing radar systems should be considered for an efficient SSA system using currently available technology. The purpose of this paper is to analyze the performance of a radar system in detecting and tracking space objects. With the radar system investigated, the minimum sensitivity is defined as detection of a $1-m^2$ radar cross section (RCS) at an altitude of 2,000 km, with operating frequencies in the L, S, C, X or Ku-band. The results of power budget analysis showed that the maximum detection range of 2,000 km, which includes the low earth orbit (LEO) environment, can be achieved with a transmission power of 900 kW, transmit and receive antenna gains of 40 dB and 43 dB, respectively, a pulse width of 2 ms, and a signal processing gain of 13.3 dB, at a frequency of 1.3 GHz. We defined the key parameters of the radar following a performance analysis of the system. This research can thus provide guidelines for the conceptual design of radar systems for national SSA initiatives.

Development of a Low-cost Industrial OCR System with an End-to-end Deep Learning Technology

  • Subedi, Bharat;Yunusov, Jahongir;Gaybulayev, Abdulaziz;Kim, Tae-Hyong
    • 대한임베디드공학회논문지
    • /
    • 제15권2호
    • /
    • pp.51-60
    • /
    • 2020
  • Optical character recognition (OCR) has been studied for decades because it is very useful in a variety of places. Nowadays, OCR's performance has improved significantly due to outstanding deep learning technology. Thus, there is an increasing demand for commercial-grade but affordable OCR systems. We have developed a low-cost, high-performance OCR system for the industry with the cheapest embedded developer kit that supports GPU acceleration. To achieve high accuracy for industrial use on limited computing resources, we chose a state-of-the-art text recognition algorithm that uses an end-to-end deep learning network as a baseline model. The model was then improved by replacing the feature extraction network with the best one suited to our conditions. Among the various candidate networks, EfficientNet-B3 has shown the best performance: excellent recognition accuracy with relatively low memory consumption. Besides, we have optimized the model written in TensorFlow's Python API using TensorFlow-TensorRT integration and TensorFlow's C++ API, respectively.