• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.3
• /
• pp.473-482
• /
• 2022

Wearing a mask is also necessary to limit the risk of infection in today's era of COVID-19 and wearing a helmet is inevitable for the safety of personnel who works in a dangerous working environment such as construction sites. This paper proposes an effective deep learning model, HelmetMask-Net, to classify both Helmet and Mask. The proposed HelmetMask-Net is based on CNN which consists of data processing, convolution layers, max pooling layers and fully connected layers with four output classifications, and 4 classes for Helmet, Mask, Helmet & Mask, and no Helmet & no Mask are classified. The proposed HelmatMask-Net has been chosen with 2 convolutional layers and AdaGrad optimizer by various simulations for accuracy, optimizer and the number of hyperparameters. Simulation results show the accuracy of 99% and the best performance compared to other models. The results of this paper would enhance the safety of personnel in this era of COVID-19.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.4
• /
• pp.577-586
• /
• 2022

Detection and classification of steel surface defects are critical for product quality control in the steel industry. However, due to its low accuracy and slow speed, the traditional approach cannot be effectively used in a production line. The current, widely used algorithm (based on deep learning) has an accuracy problem, and there are still rooms for development. This paper proposes a method of steel surface defect detection combining EfficientNetV2 for image classification and YOLOv5 as an object detector. Shorter training time and high accuracy are advantages of this model. Firstly, the image input into EfficientNetV2 model classifies defect classes and predicts probability of having defects. If the probability of having a defect is less than 0.25, the algorithm directly recognizes that the sample has no defects. Otherwise, the samples are further input into YOLOv5 to accomplish the defect detection process on the metal surface. Experiments show that proposed model has good performance on the NEU dataset with an accuracy of 98.3%. Simultaneously, the average training speed is shorter than other models.

• Journal of Dental Rehabilitation and Applied Science
• /
• v.37 no.3
• /
• pp.101-110
• /
• 2021

Purpose: In this study, we intended to study the change in bond strength according to the thermal cycling of provisional resin and 3D printed resin for making provisional restoration. Materials and Methods: Through DLP method, 3D printed resin powder was used to produce 3D printed resin samples. The samples were grouped into eight groups, according to types of provisional resin (PMMA, bis-acryl resin) which is to be bonded on the samples and numbers of thermal cycling (control, 2,000, 3,000, 5,000 cycles). Shear bond strength of the bonded samples was measured on the universal testing machine. Results: As the number of thermal cycling increased, the shear bond strength of PMMA and bis-acryl resin for 3D printed resins decreased except between 3,000 cycles and 5,000 cycles in PMMA groups. In the PMMA group, there were significant differences in shear bond strength between less number than 3,000 cycles (P < 0.05) and no significant differences between more number than 3,000 cycles (P > 0.05). In the bis-acryl resin group, there were significant differences in shear bond strength between control and 2,000 cycles, control and 3,000 cycles, and control and 5,000 cycles (P < 0.05), no significant difference between 2,000 and 3,000 cycles, between 3,000 and 5,000 cycles (P > 0.05). Conclusion: The shear bond strength between 3D printed resin and provisional resin tended to decrease after thermal cycling.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.10 no.1
• /
• pp.110-120
• /
• 1986

45.deg. corss jet flow, at the mixing of two jet flows, was experimentally studied. For this study, only the statistical turbulent characteristics and high order moments will be analysed by on-line computer system (hot-wire anemometer system, dynamic analyser and computer system, plotting and printing system). Since mean velocity distributions, intensities of turbulence, Reynolds stresses, correlation coefficients, and other general results were already studied and presented. One dimensional probability density distributions of u', v', and w' were analysed comparing with Gaussian curve, which showed skew and flat tendency according to the Y and Z directions. For the analysis of the joint flow of turublent components, the joint probability density distributions were examined. The fagures were drawn so as to be read joint probabilities, joint probability densities, fluctuating velocities u', v', and w'. For further detailed examination of the variations of skewness and flatness phenomena, iso-joint probability density contours obtained from the profiles of the joint probability density distributions were studied. According to the displacement of positions from the center of the mixing flow and the directions, the flatness and skewness factors were increased.

• Annual Conference on Human and Language Technology
• /
• 2001.10d
• /
• pp.57-62
• /
• 2001

인터넷의 급속한 양적 증가로 인해 색인어 기반의 검색 방식만으로는 원하는 정보를 찾아 내기가 쉽지 않다. 색인어 기반의 검색 방식에서는 색인어로 나타나지 않는 특징을 이용할 수 없으며, 질적으로 균등한 검색 결과를 제시하지 못하기 때문이다. 따라서 사이트의 여러 가지 특성에 따라 계층적으로 분류해놓은 웹 디렉토리를 이용하거나, 관련 전문가들의 추천 리스트를 이용하여 검객하기도 한다. 본 연구에서는 기존의 색인어 기반의 검색 모델에 웹 디렉토리와 추천 문서 같은 문서간의 링크 정보를 결합할 수 있는 정보 검색 모델을 제시한다. 특정 질의어의 검색 결과로 얻어낸 문서와 그 문서와 연결된 문서 집합을 이용하여 네트워크를 구성한다. 이 네트워크에 검색기가 제시하는 순위와 유사도, 그리고 문서간의 링크 정도를 이용해서 확률값을 정해준다. 그리고 Ergodic Markov Model의 특성을 이용하여 색인어 정보와 링크 정보를 결합한다. 본 연구에서는 특정 문서가 질의어에 부합되는 정도를 사용자가 그 문서로 이동할 확률값으로 계산하는 방식을 보인다.

• The KIPS Transactions:PartA
• /
• v.8A no.1
• /
• pp.9-15
• /
• 2001

영상에 관련된 다양한 응용 시스템들을 구현하는 많은 연구들이 진행되어 왔지만, 그러한 영상 관련 응용 시스템을 구현함에 있어서 처리속도의 저하로 인하여 많은 어려움을 겪고 있다. 이를 해결하기 위해 대두된 여러 방법들 중에서 최근 하드웨어 접근 방법에 고려한 많은 관심과 연구가 진행되고 있다. 본 논문은 영상을 실시간으로 처리하기 위하여 하드웨어 구조를 갖는 병렬처리시스템을 기술하며, 또한 병렬처리시스템을 얼굴 검색 시스템에 적용한 후 처리속도 및 실험 결과를 기술한다. 병렬처리시스템은 SIMD와 MIMD가 결합된 구조를 갖고 있기 때문에 다양한 영상 응용시스템에 대해서 융통성과 효율성을 제공하며, 144개의 처리기와 12개의 다중접근기억장치, 외부 메모리 모듈을 위한 인터페이스와 외부 프로세서 장치(i960Kx)와의 통신을 위한 인터페이스로 구성되어있다. 다중접근기억장치는 메모리 모듈선택회로, 데이터 라이팅회로, 그리고, 주소계산 및 라우팅회로로 구성되어 있다. 또한 얼굴 검색 시스템을 병렬처리 시스템에 적합한 병렬화를 제공하기 위해 메쉬방법을 이용하여 전처리, 정규화, 4개 특징값 추출, 그리고 분류화로 구성하였다. 병렬처리시스템은 하드웨어 모의실험 패키지인 CADENCE사의 Verilog-XL로 모의실험을 수행하여 기능과 성능을 검증하였다.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.8
• /
• pp.2050-2059
• /
• 1999

Multiprocessor scheduling problem has been an important research area for the past decades. The problem is defined as finding an optimal schedule which minimizes the parallel execution time of an application on a target multiprocessor system. Duplication Based Scheduling (DBS) is a relatively new approach for solving multiprocessor scheduling problems. This paper classifies DBS algorithms into two categories according to the task duplication method used. The paper then presents a new DBS algorithm that extracts the strong features of the two categories of DBS algorithms. The simulation study shows that the proposed algorithm achieves considerable performance improvement over existing DBS algorithms with similar time complexity.

• Aerospace Engineering and Technology
• /
• v.7 no.1
• /
• pp.236-244
• /
• 2008

The evolution and application of RFID technologies have been at the forefront of allowing aviation industries to improve the quality of aircraft maintenance and air cargo handling. However, safety problems in airplane operation are arising from the hazards of frequencies transmitted due to RFID systems. Though the intensities of frequencies back-scattered from the tags are very weak, some malfunctions are anticipated due to induction coupling on aircraft wiring. Therefore, safety assessment such as electromagnetic compatability should be accomplished upon aircraft critical and essential equipments before installations.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.3
• /
• pp.600-611
• /
• 1996

Recently, neural network-based speech recognition has been studied to utilize the adaptivity and learnability of neural network models. However, conventional neural network models have difficulty in the co-articulation processing and the boundary detection of similar phonmes of the Korean speech. Also, in case of using one phonotopic map, learning speed may dramatically increase and inaccuracies may be caused because homogeneous learning and recognition method should be applied for heterogenous data. Hence, in this paper, a neural net typewriter has been designed using a hierarchical self-organizing classifier(HSOC), and related algorithms are presented. This HSOC, during its learing stage, distributed phoneme data on hierarchically structured multiple phonotopic maps, using Kohonen's self-organizing feature maps(SOFM). Presented and experimented in this paper were the algorithms for deciding the number of maps, map sizes, the selection of phonemes and their placement per map, an approapriate learning and preprocessing method per map. If maps are divided according to a priorlinguistic knowledge, we would have difficulty in acquiring linguistic knowledge and how to alpply it(e.g., processing extended phonemes). Contrarily, our HSOC has an advantage that multiple phonotopic maps suitable for given input data are self-organizable. The resulting three korean phonotopic maps are optimally labelled and have their own optimal preprocessing schemes, and also confirm to the conventional linguistic knowledge.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.63 no.4
• /
• pp.534-540
• /
• 2014

In this paper, we propose a fuzzy combined Polynomial Neural Network(PNN) for pattern classification. The fuzzy combined PNN comes from the generic TSK fuzzy model with several linear polynomial as the consequent part and is the expanded version of the fuzzy model. The proposed pattern classifier has the polynomial neural networks as the consequent part, instead of the general linear polynomial. PNNs are implemented by stacking the simple polynomials dynamically. To implement one layer of PNNs, the various types of simple polynomials are used so that PNNs have flexibility and versatility. Although the structural complexity of the implemented PNNs is high, the PNNs become a high order-multi input polynomial finally. To estimate the coefficients of a polynomial neuron, The weighted linear discriminant analysis. The output of fuzzy rule system with PNNs as the consequent part is the linear combination of the output of several PNNs. To evaluate the classification ability of the proposed pattern classifier, we make some experiments with several machine learning data sets.