• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.17 no.2
• /
• pp.15-23
• /
• 2017

The person Re-identification is the most challenging part of computer vision due to the significant changes in human pose and background clutter with occlusions. The picture from non-overlapping cameras enhance the difficulty to distinguish some person from the other. To reach a better performance match, most methods use feature selection and distance metrics separately to get discriminative representations and proper distance to describe the similarity between person and kind of ignoring some significant features. This situation has encouraged us to consider a novel method to deal with this problem. In this paper, we proposed an enhanced recurrent neural network with three-tier hierarchical network for person re-identification. Specifically, the proposed recurrent neural network (RNN) model contain an iterative expectation maximum (EM) algorithm and three-tier Hierarchical network to jointly learn both the discriminative features and metrics distance. The iterative EM algorithm can fully use of the feature extraction ability of convolutional neural network (CNN) which is in series before the RNN. By unsupervised learning, the EM framework can change the labels of the patches and train larger datasets. Through the three-tier hierarchical network, the convolutional neural network, recurrent network and pooling layer can jointly be a feature extractor to better train the network. The experimental result shows that comparing with other researchers' approaches in this field, this method also can get a competitive accuracy. The influence of different component of this method will be analyzed and evaluated in the future research.

• Journal of Digital Convergence
• /
• v.12 no.5
• /
• pp.353-366
• /
• 2014

The purpose was to describe the state of healthcare-associated infection(HAI) control. Data were collected from 134 hospitals. The questionnaire developed by Kang[8] were modified. The mean of hospital beds was 556.4, 26.9% of hospitals were less than 300 beds. 99.3% of hospitals had infection control committee(ICC). ICC met 3.4 times a year. 54.5% of hospitals had one infection control practitioner(ICP). 95.5% of ICPs were nurse, 48.7% of ICPs had more than master's degree. Hospital experience of ICPs was 13.5 years. ICP experience was 3.2 years. 30.8% of ICPs worked for less than 1 year. All hospitals investigated HAI, 75.4% performed improvement activities. There are significant differences in existence of ICD, negative pressure room, computer program, numbers of ICPs according to hospital size. Manpower, organization, and facilities lacked in less than 300 beds. This conclusions will give baseline data to establish infection control system, manpower and practice in small-medium hospitals.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.42 no.6
• /
• pp.37-48
• /
• 2005

The efficiency of power supply circuits such as DC-DC converters and batteries varies on the trend of the power consumption because their efficiencies are not fixed. To analyze the efficiency of power supply circuits, we need the temporal behavior of the power consumption of the loads, which is dependent on the activity factors of the devices during the operation. Since it is not easy to model every detail of those factors, one of the most accurate power consumption analyses of power supply circuits is measurement of a real system, which is expensive and time consuming. In this paper, we introduce an active load emulator for embedded systems which is capable of power measurement, logging, replaying and synthesis. We adopt a pattern recognition technique for data compression in that long-term behaviors of power consumption consist of numbers of repetitions of short-term behaviors, and the number of short-term behaviors is generally limited to a small number. We also devise a heterogeneous structure of active load elements so that low-speed, high-current active load elements and high-speed, low-current active load elements may emulate large amount and fast changing power consumption of digital systems. For the performance evaluation of our load emulator, we demonstrate power measurement and emulation of a hard drive. As an application of our load emulator, it is used for the analysis of a DC-DC converter efficiency and for the verification of a low-power frequency scaling policy for a real-time task.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.39 no.5
• /
• pp.544-555
• /
• 2002

The aim of this study is to develop the $^1H$-MRS data postprocessing software for both single-voxel and multi-voxel technique, which plays and important role as a diagnostic tool in clinical field. This software is based on graphical user interface(GUI) under windows operating system of personal computer(PC). In case of single-voxel MRS, both of raw data in time-domain and spectrum data in frequency-domain are simultaneously displayed in a screen. Several functions such as DC correction, zero filling, line broadening, Lorentz-Gauss filtering and phase correction, etc. are included to increase the quality of spectrum data. In case of multi-voxel analysis, spectroscopic image reconstructed by 3-D FFT was displayed as a spectral grid and overlapped over previously obtained T1- or T2-weighted image for the spectra to be spatially registered with the image. The analysis of MRS peaks were performed by obtaining the ratio of peak area. In single-voxel method, statistically processed peak-area ratios of MRS data obtained from normal human brain are presented. Using multi-voxel method, MR spectroscopic image and metabolite image acquired from brain tumor are demonstrated.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.1
• /
• pp.70-80
• /
• 1999

In this paper, a new hierarchical image segmentation algorithm based on human visual system(HVS) characteristic is proposed which can efficiently reduce and control transmission information quantity without the degradation of the subjective and objective image quality. It consists of image segmentation based on mathematical morphology and region merging considering HVS characteristic for the pairs of two adjacent regions at each level of the hierarchy. Image segmentation is composed of 3-level hierarchical structure. In the region merging structure of each level, we extract the pairs of two adjacent regions which human vision can't discriminate, and then merge them. The proposed region merging method extracts pairs of two neighbor regions to be merged and performs region merging according to merging priority based on HVS characteristics. The merging priority for each adjacent pair is determined by the proposed merging priority function(MPF). First of all, the highest priority pair is merged. The information control factor is used to regulate the transmission information at each level. The proposed segmentation algorithm can efficiently improve bottleneck problem caused by excessive contour information at region-based very low bit rate coding. And it shows that it is more flexible structure than that of conventional method. In experimental results, though PSNR and the subjective image quality by the proposed algorithm is similar to that of conventional method, the contour information quantity to be transmitted is reduced considerably. Therefore it is an efficient image segmentation algorithm for region-based very low bit rate coding.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.6
• /
• pp.107-116
• /
• 2015

In compliance with digital device growth, the proliferation of high-tech computers, the availability of high quality and inexpensive video cameras, the demands for automated video analysis is increasing, especially in field of intelligent monitor system, video compression and robot vision. That is why object tracking of computer vision comes into the spotlight. Tracking is the process of locating a moving object over time using a camera. The consideration of object's scale, rotation and shape deformation is the most important thing in robust object tracking. In this paper, we propose a robust object tracking scheme using Random Forest. Specifically, an object detection scheme based on region covariance and ZNCC(zeros mean normalized cross correlation) is adopted for estimating accurate object location. Next, the detected region will be divided into five regions for random forest-based learning. The five regions are verified by random forest. The verified regions are put into the model pool. Finally, the input model is updated for the object location correction when the region does not contain the object. The experiments shows that the proposed method produces better accurate performance with respect to object location than the existing methods.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.48 no.1
• /
• pp.1-7
• /
• 2011

In this paper, we propose the digital image enhancement method including local tone reproduction and preservation of the hue. In recent studies, an integrated multi-scale retinex (IMSR) has produced great naturalness in the resulting images through enhancement of visibility in dark area in input images. However, most methods, including IMSR, work in RGB color spaces. As such, this produces hue distortion from the perspective of the human visual system, that is, hue distortion in CIELAB color space. Accordingly, this paper proposes an tone reproduction and enhancement of saturation method in a device-independent color space, CIELAB, to preserve the hue and obtain a high contrast and naturalness. First, to achieve the desired objectives, the IMSR is then applied to only the $L^*$ values in CIELAB color space, normalization, and simple mapping function, thereby preserving the balance of the color components and enhancement of visibility. Then, saturation adjustment is performed by applying the ratio of the chroma variation at the sRGB gamut boundary according to the corrected luminance. In experiments, the proposed method is shown to improve the visibility in dark shadows and bright regions in the resulting images and reduce any color distortion then preference test are performed.

• Journal of the Korea Concrete Institute
• /
• v.17 no.6 s.90
• /
• pp.955-962
• /
• 2005

Still no accurate theory exists for predicting ultimate shear strength of deep reinforced concrete beams because of the structural and material non-linearity after cracking. Currently, the load capacity assesment is performed for the upper structure of the bridges and containing non-reliability in the applications and results. The purpose in this study is to evaluate analytically the complex shear behaviors and normal strength for the reinforced concrete deep beams and to offer the accuracy load capacity assesment method based on the reliability theories. This paper presents a method for the load capacity assesment of reinforcement concrete deep beams using nonlinear finite element analysis. A computer program named RCAHEST (Reinforced Concrete Analysis in Higher Evaluation System Technology), for the analysis of reinforced concrete structures was used. Material non-linearity is taken Into account by comprising tensile, compressive and shear models of cracked concrete and a model of reinforcing steel. From the results, determine the reliability index for the failure base on the Euro Code. Then, calculate additional reduction coefficient to satisfy the goals from the reliability analysis. The proposed numerical method for the load capacity assesment of reinforced concrete deep beams is verified by comparison with the others methods.

• Journal of Korean Academy of Nursing
• /
• v.20 no.1
• /
• pp.61-78
• /
• 1990

The purpose of the present study is to confirm the interacting behavior between nurses and patients and other things concerned herewith. Subjects of investigation were : 42 nurses selected out of the average nurses who serve in hospital as nurses assigned to medical and surgical wards : and 42 male and female adult patients selected out of the average patients who were under the care of the nurse individuals and can make themselves understood verbally. A nurse and her patient were paired off for questioning. Materials for statistics were gathered by means of observaing interactions- - verbal and nonverbal - -of the chosen subjects for four hours every day from 7 : 30 a.m. through 7 : 30 p.m. between on July 15, 1988 and on Aug. 16, 1988. Classified by patterns, the materials observed and gathered were preliminarily analyzed by this researcher, and then reexamined in a full-fledged way by one professor, three nurses and three non - nurses. The researcher depended chiefly on Frequency, ANOVA, Pearson's Correlation Coefficient attached to SPSS Computer System for the process of gathered materials. The results of this investigations are follows 1) A total of 98 times' interactions between nurses and patients were provided during observation of 168 hours. 2) It took them the averaged 264.8 seconds(around 4.4 minutes) per a couple of subjects to interact between nurses and patients during observation of four hours. 3) The aim of interactions between nurses and patients appear that 29 times of injection amounted to 29.6% the most, 27 times of PO around to 27.6% the next most, 25 times of vital check to 25.5% the next most, 17 times of independent nursing works and round to 17.3% the least most. 4) As a result of qualitative analyzing the interactions between nurses and patients by the distinctive method of words were positively recognized in 19 cases with 45.2% and negatively in 23 cases with 54.8%. 5) A total of 2, 193 times. interaction behaviours between nurses and patients were provided. The frequency of these interaction behaviours took place l, 364 times with 62.2% to nurse, and 829 times with 37.8% to patients. 6) The classification of verbal and nonverbal interaction behaviour between nurses and patients indicated that it is amounted to 64.9% for verbal behaviour numbered 1, 423 and 35.1% for nonverbal one numbered 770. 7) The frequency of verbal behaviour between nurses and patients numbered 1, 423 in total. They took place 924 times to nurses and 499 times to patients, it can be also amounted to 64.9% and 35.1% respectively in percentagewise. 8) In interactions between nurses and patients, it turned out that the frequency of nurses' turns, which the present research discovered averaged 16.8 times for four hours, and the verbal behaviours by numbered 9.7 on an average. 9) Nonverbal behaviours between nurses and patients numbered 770 in total, it is assigned 440 times to nurse with 57.1% and 330 times to patients with 42.9%. 10) The investigation releases in formation that the frequency of verbal behaviours between nurses and patients was very much concerned with the age of patients(r＝0.422, p＜.01) and the number of patients one nurse has under her care(r＝-0.356, p＜.01). 11) It was found that were deep relationship of the number of a nurses turn with the patients age(r＝0.377, p＜.01) and the nurses burden of caring patients(r＝-0.372, p＜.01).

• Imaging Science in Dentistry
• /
• v.33 no.2
• /
• pp.113-119
• /
• 2003

Purpose : To evaluate the relationship between morphometric analysis of bone microstructure from digital radiographic image and trabecular bone strength. Materials and Methods : One hundred eleven bone specimens with 5 mm thickness were obtained from the mandibles of 5 pigs. Digital images of specimens were taken using a direct digital intraoral radiographic system. After selection of ROI (100 × 100 pixel) within the trabecular bone, mean gray level and standard deviation were obtained. Fractal dimension and the variants of morphometric analysis (trabecular area, periphery, length of skeletonized trabeculae, number of terminal point, number of branch point) were obtained from ROI. Punch sheer strength analysis was performed using Instron (model 4465, Instron Corp., USA). The loading force (loading speed 1 mm/min) was applied to ROI of bone specimen by a 2 mm diameter punch. Stress-deformation curve was obtained from the punch sheer strength analysis and maximum stress, yield stress, Young's modulus were measured. Results: Maximum stress had a negative linear correlation with mean gray level and fractal dimension significantly (p<0.05). Yield stress had a negative linear correlation with mean gray level, periphery, fractal dimension and the length of skeletonized trabeculae significantly (p < 0.05). Young's modulus had a negative linear correlation with mean gray level and fractal dimension significantly (p < 0.05). Conclusions : The strength of cancellous bone exhibited a significantly linear relationship between mean gray level, fractal dimension and morphometric analysis. The methods described above can be easily used to evaluate bone quality clinically.