• Journal of IKEEE
• /
• v.21 no.4
• /
• pp.388-396
• /
• 2017

Since there are many variables such as various poses, illuminations and occlusions in a face detection problem, a high performance detection system is required. Although CNN is excellent in image classification, CNN operatioin requires high-performance hardware resources. But low cost low power environments are essential for small and mobile systems. So in this paper, the CPU-FPGA integrated system is designed based on 3-stage cascade CNN architecture using small size FPGA. Adaptive Region of Interest (ROI) is applied to reduce the number of CNN operations using face information of the previous frame. We use a Field Programmable Gate Array(FPGA) to accelerate the CNN computations. The accelerator reads multiple featuremap at once on the FPGA and performs a Multiply-Accumulate (MAC) operation in parallel for convolution operation. The system is implemented on Altera Cyclone V FPGA in which ARM Cortex A-9 and on-chip SRAM are embedded. The system runs at 30FPS with HD resolution input images. The CPU-FPGA integrated system showed 8.5 times of the power efficiency compared to systems using CPU only.

• Journal of Intelligence and Information Systems
• /
• v.14 no.1
• /
• pp.1-17
• /
• 2008

Data mining is defined as the process of discovering meaningful and useful pattern in large volumes of data. In particular, finding associations rules between items in a database of customer transactions has become an important thing. Some data structures and algorithms had been proposed for storing meaningful information compressed from an original database to find frequent itemsets since Apriori algorithm. Though existing method find all association rules, we must have a lot of process to analyze association rules because there are too many rules. In this paper, we propose a new data structure, called a Frequent Pattern Network (FPN), which represents items as vertices and 2-itemsets as edges of the network. In order to utilize FPN, We constitute FPN using item's frequency. And then we use a clustering method to group the vertices on the network into clusters so that the intracluster similarity is maximized and the intercluster similarity is minimized. We generate association rules based on clusters. Our experiments showed accuracy of clustering items on the network using confidence, correlation and edge weight similarity methods. And We generated association rules using clusters and compare traditional and our method. From the results, the confidence similarity had a strong influence than others on the frequent pattern network. And FPN had a flexibility to minimum support value.

• Korean Journal of Materials Research
• /
• v.9 no.12
• /
• pp.1263-1269
• /
• 1999

An FBAR(Solidly Mounted Resonator) was fabricated using reflector layers which prohibit the penetration of bulk acoustic wave into substrate. The SMR consisted of top and bottom electrodes(Al films), a piezoelectric layer (ZnO film), reflector layers(W/$Si_2$ films) and Si substrate. The electrodes were deposited by dc sputtering. The piezoelectric layer and the reflector layers were deposited by rf magnetron sputtering. The control of crystallinity, microstructures and electric properties of each layer was essential for attaining the optimum FBAR characteristics. Under the best deposition conditions for FBAR devices, the ZnO films had highly c-axis preferred orientation(${\sigma}=2.17^{\circ}$), resistivity of $10^4\;{\omega}cm$, and surface roughness of 10.6 ${\AA}$. On the other hand, the surface roughness of W and $Si_2$ films was 16 ${\AA}$ and 33 ${\AA}$, respectively, and the resistivity of Al film was $5.1{\times}10^{-6}\;{\Omega}cm$. The SMR devices were fabricated by the conventional semiconductor processes. In the resonance conditions of the SMR, the series resonance frequency (fs) and the parallel resonance frequency(fp) were 1.244 GHz and 1.251 GHz, respectively and the quality factor(Q) was 1200.

• Journal of the Korean Society of Radiology
• /
• v.7 no.4
• /
• pp.277-284
• /
• 2013

This study aims to find out geometric parameters which practitioner adjustable to reduce dose in coronary angiography. We take fluoroscopy and cine exposure by use of phantom, and got dose use the dose-area product(DAP) meter of angiography device, than convert DAP to effective dose. As results, Cine exposure shows higher dose measurement about 6-7 times than fluoroscopy. Dose in frame per second(FPS) mode could be decrease down to 70%, as lower FPS. In view of X-ray tube angle, LAO $45^{\circ}$+Caudal $30^{\circ}$ shows highest dose measurement. More use of Collimator, lower dose measurement. Source-image intensifier distance(SID) get longer to 10cm, dose of each fluoroscopy and cine exposure increase up to 25-30%. Image magnification of field of view(FOV) could increase dose up to 1.21-2 times. Also table-image intensifier distance get longer to 10cm, dose increased 1.11-1.25 times. Practitioner can adjust several geometric parameters, as FPS mode, tube angle, Collimation, SID, table-image intensifier distance, FOV. And each factors can reduce radiation dose in coronary angiography.

• Journal of Distribution Science
• /
• v.9 no.3
• /
• pp.55-64
• /
• 2011

This study is important in its focus to find key clues in the marketing strategy, consumer behavior, and communication processes that define the infantile obesity market. The study, the first of its kind, surveyed a target audience, purchasing group, and housewives in their quest to determine purchasing decisions and effective branding strategy planning for the infantile obesity market. Another key component of the study was to focus on the key direct and/or indirect distribution channels for the subject market. Recently, obesity has emerged as a major social concern; some studies show that the onslaught of an adverse eating culture in Korea emanates from the prevalence of fast-food dining establishments. Obesity among children leads to adult obesity, especially if the young people's parents are overweight; notably, if either one or both of the parents are obese, the percentage of young people eventually being obese is approximately 80 to 85 percent. Because obesity is the cause of many major health concerns later in life, the struggle for a healthy life is considerably adversely affected by parents' consumer behavior. Infantile obesity, resulting in adult obesity, is also an important national economic and social issue. The sizable direct and indirect economic costs, as well as the tremendous social costs of obesity, cannot be overstated. Effective food branding and advertising centered on food preferences and dietary behaviors, especially to children, creates an effective marketing effort that, ultimately, leads to positive results. Thus, the purpose of this study is to demonstrate that the treatment of childhood obesity in Korea, through the activation of a brand and retail market, can effectively solve social and economic problems that result from infantile and childhood obesity. In this study, obesity markets and distribution channels in the purchase decision-making factors determining factor based on it effective inspection and branding strategies and brand marketing communications strategy proposed measures contribute to the obesity drug market and further enable the childhood obesity problem is intended to assist in solving.

• Journal of Bio-Environment Control
• /
• v.30 no.4
• /
• pp.393-400
• /
• 2021

Body weight of livestock is a crucial indicator for assessing feed requirements and nutritional status. This study was performed to estimate the body weight of Korean cattle (Hanwoo) using body volume determined from three-dimensional (3-D) image. A TOF camera with a resolution of 640×480 pixels, a frame rate of 44 fps and a field of view of 47°(H)×37°(V) was used to capture the 3-D images for Hanwoo. A grid image of the body was obtained through preprocessing such as separating the body from background and removing outliers from the obtained 3-D image. The body volume was determined by numerical integration using depth information to individual grid. The coefficient of determination for a linear regression model of body weight and body volume for calibration dataset was 0.8725. On the other hand, the coefficient of determination was 0.9083 in a multiple regression model for estimating body weight, in which the age of Hanwoo was added to the body volume as an explanatory variable. Mean absolute percentage error and root mean square error in the multiple regression model to estimate the body weight for validation dataset were 8.2% and 24.5kg, respectively. The performance of the regression model for weight estimation was improved and the effort required for estimating body weight could be reduced as the body volume of Hanwoo was used. From these results obtained, it was concluded that the body volume determined from 3-D of Hanwoo could be used as an effective variable for estimating body weight.

• Journal of Korean Library and Information Science Society
• /
• v.50 no.1
• /
• pp.23-41
• /
• 2019

This study understand the subject domain of book which has been read by men and woman elderly by analizying the PFNET using library big data and confirm the difference between adult at age 30-40. This study extract co-occurrence matrix of book lending on the popular book list from library big data, for 4 group, men/woman elderly, men/woman adult. With these matrix, this study performs FP network analysis. And Pearson Correlation Analysis based on the Triangle Betweenness Centrality calculated on the loan book was performed to understand the correlation among the 4 clusters which has been created by PNNC algorithm. As a result, reading trend which has been focused on modern korean novel has been revealed in elderly regardless gender, among them, men elderly show extreme tendency concentrated on modern korean long series novel. In the correlation analysis, the male elderly showed a weak negative correlation with the adult male of r = -0.222, and the negative direction of all the other groups showed that the tendency of male elderly's loan book was opposite.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.18 no.4
• /
• pp.58-70
• /
• 2019

In the case of a pedestrian traffic accident, it has a large-scale danger directly connected by a fatal accident at the time of the accident. The domestic ITS is not used for intelligent risk classification because it is used only for collecting traffic information despite of the construction of good quality traffic infrastructure. The CNN based pedestrian detection classification model, which is a major component of the proposed system, is implemented on an embedded system assuming that it is installed and operated in a restricted environment. A new model was created by improving YOLO's artificial neural network, and the real-time detection speed result of average accuracy 86.29% and 21.1 fps was shown with 20,000 iterative learning. And we constructed a protocol interworking scenario and implementation of a system that can connect with the ITS. If a pedestrian accident prevention system connected with ITS will be implemented through this study, it will help to reduce the cost of constructing a new infrastructure and reduce the incidence of traffic accidents for pedestrians, and we can also reduce the cost for system monitoring.

• Journal of Broadcast Engineering
• /
• v.24 no.3
• /
• pp.495-505
• /
• 2019

Typical algorithms of deep learning include CNN(Convolutional Neural Networks), which are mainly used for image recognition, and RNN(Recurrent Neural Networks), which are used mainly for speech recognition and natural language processing. Among them, CNN is able to learn from filters that generate feature maps with algorithms that automatically learn features from data, making it mainstream with excellent performance in image recognition. Since then, various algorithms such as R-CNN and others have appeared in object detection to improve performance of CNN, and algorithms such as YOLO(You Only Look Once) and SSD(Single Shot Multi-box Detector) have been proposed recently. However, since these deep learning-based detection algorithms determine the success of the detection in the still images, stable object tracking and detection in the video requires separate tracking capabilities. Therefore, this paper proposes a method of combining Kalman filters into deep learning-based detection networks for improved object tracking and detection performance in the video. The detection network used YOLO v2, which is capable of real-time processing, and the proposed method resulted in 7.7% IoU performance improvement over the existing YOLO v2 network and 20 fps processing speed in FHD images.

• Journal of Broadcast Engineering
• /
• v.24 no.1
• /
• pp.132-141
• /
• 2019

The necessity of transmitting video data over a narrow-bandwidth exists steadily despite that video service over broadband is common. In this paper, we propose a scalable video coding framework for low-resolution video transmission over a very narrow-bandwidth network by super-resolution of decoded frames of a base layer using a convolutional neural network based super resolution technique to improve the coding efficiency by using it as a prediction for the enhancement layer. In contrast to the conventional scalable high efficiency video coding (SHVC) standard, in which upscaling is performed with a fixed filter, we propose a scalable video coding framework that replaces the existing fixed up-scaling filter by using the trained convolutional neural network for super-resolution. For this, we proposed a neural network structure with skip connection and residual learning technique and trained it according to the application scenario of the video coding framework. For the application scenario where a video whose resolution is $352{\times}288$ and frame rate is 8fps is encoded at 110kbps, the quality of the proposed scalable video coding framework is higher than that of the SHVC framework.