• 한국멀티미디어학회논문지
• /
• 제23권7호
• /
• pp.803-811
• /
• 2020

This paper proposes an image classification algorithm that transforms the number of convolution layers in the residual block of ResNet, CNN's representative method. The proposed method modified the structure of 34/50 layer of ResNet structure. First, we analyzed the performance of small and many convolution layers for the structure consisting of only shortcut and 3 × 3 convolution layers for 34 and 50 layers. And then the performance was analyzed in the case of small and many cases of convolutional layers for the bottleneck structure of 50 layers. By applying the results, the best classification method in the residual block was applied to construct a 34-layer simple structure and a 50-layer bottleneck image classification model. To evaluate the performance of the proposed image classification model, the results were analyzed by applying to the cifar10 dataset. The proposed 34-layer simple structure and 50-layer bottleneck showed improved performance over the ResNet-110 and Densnet-40 models.

• International Journal of Internet, Broadcasting and Communication
• /
• 제12권3호
• /
• pp.220-225
• /
• 2020

This paper investigates the non-intrusive speech intelligibility estimation method in noise environments when the bottleneck feature of autoencoder is used as an input to a neural network. The bottleneck feature-based method has the problem of severe performance degradation when the noise environment is changed. In order to overcome this problem, we propose a novel non-intrusive speech intelligibility estimation method that adds the noise environment information along with bottleneck feature to the input of long short-term memory (LSTM) neural network whose output is a short-time objective intelligence (STOI) score that is a standard tool for measuring intrusive speech intelligibility with reference speech signals. From the experiments in various noise environments, the proposed method showed improved performance when the noise environment is same. In particular, the performance was significant improved compared to that of the conventional methods in different environments. Therefore, we can conclude that the method proposed in this paper can be successfully used for estimating non-intrusive speech intelligibility in various noise environments.

• 한국조리학회지
• /
• 제13권1호
• /
• pp.179-191
• /
• 2007

This paper aims to improve management and make the bakery business stable by investigating the difficulties in bakery management and the current management conditions of bakery businesses. Using Windows SPSS 11.0, the bottleneck in bakery management and its influence on sales and job satisfaction were analyzed. Besides, a regression analysis was conducted to investigate the effect of difficulties in bakery management on sales and job satisfaction. According to the analysis, a total of 44 variables were observed as the difficulties in bakery management. Among them, 34 variables were caused by 7 factors: facility, product, employee, finance, production, external advertisement. In the regression analysis, it has turned out that the difficulties in management had influence on sales dissatisfaction in facility, product, employee, finance, production, and external factors. Especially, employee factor turned out the most influential one on sales dissatisfaction. Furthermore, it has turned out that the bottleneck in management had influence on job dissatisfaction in all 7 factors. Above all, the external factor turned out the most influential one.

• 공학교육연구
• /
• 제24권3호
• /
• pp.58-69
• /
• 2021

It is very necessary to have a creative problem-solving capacities to learn various majors and liberal arts based on the major, and to solve the bottleneck techniques led by students. In this study, the existing creative problem-solving curriculums, 'Methodology of Inventive Problem Solving' based on TRIZ, were improved and applied, and industrial bottleneck techniques were provided to students to solve these techniques. To improve the curriculum, 1) improvement of instructional objectives and learning contents, 2) improvement of evaluation methods and contents (reflecting the evaluation of instructor and students), and 3) learning satisfaction survey were conducted in the following order. As a result of the application of the improved curriculum, the level of activities for each team was improved, and when the core process was well understood, the evaluation of team activities was also excellent, but there was a tendency to focus on methods that are relatively easy to apply in the problem solving process. In the final exam (learning contents evaluation), teams with difficult understanding of the TRIZ theory or low team activities showed a relatively high trend, but the difference in level between divisions was slightly reduced.

• 한국정보통신학회논문지
• /
• 제25권12호
• /
• pp.1835-1845
• /
• 2021

본 논문에서는 블랙 아이스를 정확하게 인식하고 도로 노면 정보를 운전자에게 미리 알려줘서 속도를 제어하고 예방 조치를 취할 수 있도록 하기 위해 열화 도로 영상을 기반으로 블랙 아이스 검출하기 위해 lightweight 네트워크를 제안한다. 전이학습을 이용하여 블랙 아이스 인식 실험을 하였고, 블랙 아이스 인식의 정확도 향상을 위해 MobileNetV2 기반의 개선된 lightweight 네트워크를 개발하였다. 계산량을 줄이기 위해 Linear Bottleneck 및 Inverted Residuals를 활용하여 4개의 Bottleneck 그룹을 사용하고 모델의 인식률 향상을 위해 각 Bottleneck 그룹에 3×3 컨볼루션 레이어를 연결하여 지역적 특징 추출을 강화하고 특징 맵의 수를 늘렸다. 마지막으로 구축된 블랙 아이스 데이터 세트 대상으로 블랙 아이스 인식 실험을 진행하였으며, 제안된 모델은 블랙 아이스에 대해 99.07%의 정확한 인식률을 나타내었다.

• ETRI Journal
• /
• 제22권4호
• /
• pp.1-12
• /
• 2000

Designing a reliable multicast-based network that scales to the size of a multicast group member is difficult because of the diversity of user demands. The loss inferences of internal nodes by end-to-end measurements do not require the use of complete statistics because of the use of maximum likelihood estimation. These schemes are very efficient and the inferred value converges fast to its true value. In the theoretical analysis, internal delay estimation is possible but the analysis is very complex due to the continuity property of the delay. In this paper, we propose the use of a bottleneck location estimator. This can overcome the analytical difficulty of the delay estimation using the power spectrum of the packet interarrival time as the performance metric. Both theoretical analysis and simulation results show that the proposed scheme can be used for bottleneck location inference of internal links in scalable multicasting.

• 한국게임학회 논문지
• /
• 제2권1호
• /
• pp.30-35
• /
• 2002

The deployment of multicast communication services in the Internet is expected to lead a stable packet transfer even in heavy traffic as in On-Line Game environment. The Core Based Tree scheme among many multicast protocols is the most popular and suggested recently. However, CBT exhibit two major deficiencies such as traffic concentration or poor core placement problem. So, measuring the bottleneck link bandwidth along a path is important for understanding the performance of multicast. We propose not only a definition of CBT's core link state that Steady-State(SS), Normal-State(NS) and Bottleneck State(BS) according to the estimation link speed rate, but also the changeover of multicast routing scheme for traffic overload. In addition, we introduce anycast routing tree, a efficient architecture for construct shard multicast trees.

• 디지털산업정보학회논문지
• /
• 제5권2호
• /
• pp.101-107
• /
• 2009

In existing wireless ad-hoc networks, how to distribute network resources fairly between many users to optimize data transmission is an important research subject. However, in wireless mesh networks (WMNs), it is one of the research areas to avoid gateway bottleneck more than the fair network resource sharing. It is because WMN traffic are concentrated on the gateway connected to backhaul. To solve this problem, the paper proposes Weighted Fairness Time-sharing Access (WFTA). The proposed WFTA is a priority time scheduling scheme based on Weighted Fair Queuing (WFQ).

• Journal of Korean Vacuum Science & Technology
• /
• 제4권1호
• /
• pp.27-32
• /
• 2000

We have studied the carrier relaxation of InAs/GaAs modulation-doped quantum dots depending on the excitation wavelength and modulation-doping concentration by using the time-ressolved spectroscopy. At the excitation below GaAs barrier band gap, the relaxation processes become very slow, implying to observe the phonon bottleneck effects. On the other hand, at the excitation far above GaAs band gap, phonon bottleneck effects are broken down due to Auger processes. Increasing modulation-doping concentration, the relaxation times, by virtue of Coulomb scattering between electrons in GaAs doped layer and carriers in InAs quantum dots, are observed to become fast.

• Industrial Engineering and Management Systems
• /
• 제6권1호
• /
• pp.83-93
• /
• 2007

This paper gives an attempt to build a decision support tool linked with a simulation software called ARENA for evaluating and comparing the performance of the push and pull material driven strategies operating in the flow shop environment with a bottleneck resource as the shop's constraint. To be fair for such evaluation, the comparison must be made fairly under the optimal setting of both systems' operating parameters. In this study, an optimal-seeking heuristic algorithm, Genetic Algorithm (GA), is employed to suggest a systems' best design based on the economic consideration, which is the profit generated from the system. Results from the study have revealed interesting outcomes, letting us know the strength and weakness of the push and pull mechanisms as well as the effect of each operating parameter to the overall system's financial performance.