Earticle

Our objective was to evaluate the CT attenuation coefficient and noise of spatial domain filtering as an alternative to additional image reconstruction using different kernels in abdominal CT. Derived from thin collimated source images was generated using abdomen BIO (very smooth), B20 (smooth), B30 (medium smooth), B40 (medium), B50 (medium sharp), B60 (sharp), B70 (very sharp) and B80 (ultra sharp) kernels. Quantitative CT coefficient and noise measurements provided comparable HU (hounsfield) units in this respect. CT attenuation coefficient (mean HU) values in the abdominal were 60.4 〜 62.2 HU and noise (7.6 〜 63.8 HU) in the liver parenchyma. In the stomach a mean (CT attenuation coefficient) of -2.2 - 0.8 HU and noise (10.1 〜 82.4 HU) was measured. Image reconstructed with a convolution kernel led to an increase in noise, whereas the results for CT attenuation coefficient were comparable. Image medications of image sharpness and noise eliminate the need for reconstruction using different kernels in the future. CT images increase the diagnostic accuracy may be controlled by adjusting CT various kernels, which should be adjusted to take into account the kernels of the CT undeigoing the examination.

The objective of the present experiment was to investigate the effect of dietary palm kernel meal (PKM) and β-xylanase supplementation on productive performance, egg quality, fatty liver incidence, and excreta characteristics in laying hens. A total of 320 Hy-Line Brown laying hens (33 weeks of age) were allotted to 1 of 4 treatments with 8 replicates in a feeding trial. Each replicate consisted of 10 consecutive cages with 1 hen per cage. The corn-soybean meal-based control diet was prepared. Additional diet was prepared by including 10% of PKM in the control diet with a partial replacement of corn, soybean meal, and animal fat. In addition, 0.025% β-xylanase was supplemented at the expense of celite to those 2 diets to produce 4 treatment diets in a 2 × 2 factorial arrangement. All hens were provided the diet and water ad libitum for 8 weeks. Results indicated no significant interactions between inclusion of dietary PKM and β-xylanase for all measurements; therefore, the main effects were mainly discussed. Hens fed diets containing 10% PKM had greater (p < 0.05) feed intake and yolk color than those fed diets containing no PKM. However, dietary PKM did not influence fatty liver incidence and excreta characteristics. Dietary β-xylanase supplementation had no effects on all measurements, regardless of inclusion of PKM. In conclusion, PKM can be a potential feed ingredient for laying hens at the inclusion of 10% in the diet. It appears that dietary β-xylanase used in the current experiment has little effect on layer productivity, regardless of inclusion of 10% PKM in the diet.

This study was conducted to evaluate the effect of dietary β-mannanase supplementation and palm kernel meal(PKM) inclusion (5%) on laying performance, egg quality and nutrient utilizability of laying hens with 73 weeks of age. A total of 240 Lohmann brown laying hens with average 77.5% egg production were randomly allocated with 60 hens per treatment, 4 replicates per treatment and 15 hens per replicate. Experimental design was a completely randomized design with 2×2 factorial arrangement, with the factors being (1) two levels of PKM(0 vs. 5%) and (2) with or without dietary β-mannanase(480IU/kg of diet CTCzyme®) supplementation. All hens were housed in cages(35cmW×35cmD×40cmH) with 2 hens per cage for six weeks feeding trial. Laying performance was recorded daily during feeding trial. Egg quality, nutrients utilizability and blood assays were done at the end of feeding trial. Egg production was improved(P<0.05) by both dietary PKM inclusion and β-mannanase combined supplementation. Either β-mannanase or PKM did not affect feed intakes and feed conversion ratio of all diets. Egg weight of hens fed diet containing 5% of PKM had heavier(P<0.05) eggs compared with hens fed without PKM. Albumen height was improved(P<0.05) by dietary mannanase supplementation. Crude fat utilization of 5% PKM diet was higher than that of no PKM diet regardless of β-mannanase supplementation. Both DM and total carbohydrate utilization were decreased(P<0.05) in hens fed 5% PKM diet. Serum IgG and yolk IgY contents of PKM groups were lower(P<0.05) than those of no PKM groups. This result showed that 5% PKM diet, independent of dietary β-mannanase supplementation, was able to improve egg production. In addition, dietary β-mannanase supplementation could be used for improving the albumen height of eggs.

RF 기반 고속 무선통신 기술이 급속히 발전함에 따라 무선 주파수 대역을 기반으로한 IoT 네트워크용 디바이스가 빠르게 보급되고 있으나, 최근 IoT 네트워크 디바이스의 급속한 확산 속도에 비하여 RF 통신 기술의 발전속도가 미치지 못하고 있다. 이러한 상황에서 가시광원을 송신수단으로 사용하는 OWC 기술은 RF 기반 무선 통신의 대역 고갈 문제를 극복 할 수 있는 기술로서 주목받고 있으나, OWC용 데이터 수신 중 발생하는 LED 조명 형태의 왜곡으로 인하 여 LED 조명 검출율이 저하되고 RoI의 설정이 부정확해지는 현상이 발생 할 가능성이 있다. 본 논문에서는 Adaptive Median Filter를 적용한 저속 카메라 통신용 LED 조명 검출 알고리즘을 제안하였다. 이를 통해 명확한 RoI 설정 및 LED 조명 검출이 가능할 것으로 사료되며, 본 연구 결과를 통해 RF 기반 무선 통신기술의 보완재로서의 역할을 효율적으로 수행 할 수 있을 것으로 판단된다.

With the rapid development of RF based high speed wireless communication technology, devices that can be applied to IoT networks based on RF bandwidth are rapidly spreading, nevertheless, the development speed of the RF communication is not possible to keep up with the spread of the RF band for wireless communication. In this situation, OWC technology that uses visible light source as a transmitter is attracting attention as a technology that can overcome the band exhaustion problem of RF based wireless communication technology. Although, due to the distortion of the LED illumination shape by camera exposure time and LED blinking period, the LED illumination detection rate is degraded and the RoI setting is inaccurate. In this paper, we propose an adaptive median filter applied LED illumination detection algorithm for low rate CamCom, it is possible to detect a clear RoI and LED illumination. This research will be able to play a role as a complementary material of RF based wireless communication technology efficiently.

본 논문에서는 커널 Extreme Learning Machine을 기반으로 하여 최적화 기법들 중의 하나인 입자 군집 최적화 기법을 이용한 설계 기법을 제시한다. 제안된 Kernel Extreme Learning Machine은 기존에 사용되어지는 뉴럴 네트워크의 단점을 개선한 네트워크이다. 다시 말하면, 뉴럴 네트워크의 단점인 느린 학습속도를 개선한 네 트워크이다. 일반적으로 뉴럴 네트워크의 히든 노드들은 랜덤 초기화 후 오류 역전파 알고리즘을 이용하여 학습 한다. 이와 같은 오류 역전파 알고리즘은 매우 느린 학습속도를 보인다. 이와 같은 단점을 해결하기 위하여, Kernel Extreme Learning Machine의 히든 노드들은 랜덤 초기화 되고 학습되지 않고 출력층의 연결 하중만 학습 되어진다. 이와 같은 장점을 가진 Kernel Extreme Learning Machine의 구조를 최적화하기 위하여 입자 군집 최적 화 기법을 사용한다. 제안된 설계 방법을 적용하여 설계된 모델의 일반화 성능의 우수성을 보이기 위하여, 다수 의 머신러닝 데이터들을 이용하여 실험하고 실험을 통해 얻은 결과를 비교 평가하였다.

In this paper, we proposed the design method of Extreme Learning Machine which is optimized by using Particle Swarm Optimization Technique. Extreme Learning Machine is the improved version of the conventional neural networks which have a very slow learning speed based on the back-propagation algorithm. In the conventional neural networks, the connection weights between the input layer and the hidden layers are initialized randomly and then optimized by using the gradient decent method. The speed of the learning method for the conventional neural networks is slow. In order to overcome the drawback of the conventional neural networks, the connection weights of the hidden nodes are just initialized randomly and will not be optimized, and the only connection weights of the output nodes are learned by using least square estimation not the iterative learning method. In addition, we use Particle Swarm Optimization to optimize the proposed Extreme Learning Machine. Several machine learning bench-mark data sets are used to show the generalization performance of the proposed design method and to compare their performance with the other already studied models.

주택가격 예측은 개인, 기관, 정부에 이르기까지 다양한 이해관계자들이 관심을 갖는 주요한 주 제이다. 지금까지 통계적 모델을 통해 예측을 해왔지만, 최근 딥러닝 및 머신러닝을 이용하여 예측 하는 연구가 많다. 특히 이미지 정보를 활용한 합성곱신경망(Convolutional Neural Networks)을 적 용한 연구도 진행되고 있다. 그러나 아직 국내에서는 이미지 정보를 이용한 연구는 살펴볼 수 없었 다. 본 연구는 커널밀도추정(Kernel Density Estimation)을 활용하여 인공위성 이미지와 상가 및 편 의시설 밀도 정보를 해당 구역에 표시하고 이를 입력변수로 활용하여 서울시 아파트 가격을 예측하 고자 한다. 본 논문에서는 2021년부터 2022년까지의 주택 거래 기록과 지역소득정보, 인구 밀도, 연령대별 인구 데이터를 수집하였고, 네이버 클라우드 플랫폼의 Static Map API를 통해 인공위성사진을 구역 별로 나누어서 수집했다. 특히 공공데이터포털과 서울 열린데이터 광장에서 수집한 상가, 병원, 공 원, 지하철역, 학교 등 정보는 커널밀도추정을 활용해 밀도 정보로 변환하였다. 이렇게 다양한 형태 로 수집된 데이터는 전처리 과정을 거쳐 면적당 단가로 계산된 아파트 실거래가를 예측하였다. 예 측모델로서는 회귀 모델, 다층 인공신경망 모델, 그리고 합성곱신경망 모델을 이용하였고 이들의 예 측력을 비교하였다. 연구결과는 다음과 같다. 첫째, 회귀 모델 중에서는 인구통계 및 주택관련 변수 에 커널밀도추정으로 구해진 상가 및 편의시설 밀도 정보를 추가한 모델이 예측력이 높았다. 둘째, 다층인공신경망(Multilayer Artificial Neural Network) 모델은 회귀 모델에 비해 상대적으로 높은 성능을 보였다. 셋째, 합성곱신경망과 다층인공신경망을 융합한 모델은 인공위성 이미지 피처, 인구 통계 및 주택 관련 변수, 그리고 커널밀도추정 특성을 모두 사용한 경우 기존 다층 인공신경망과 회귀모델과 비교시 가장 우수한 성능을 보였다. 이 결과는 인공위성 이미지가 주택 가격 예측에 유 용한 정보를 제공하는 것을 시사하며 향후 주택가격예측에 인공위성 이미지, 혹은 커널밀도추정으 로 표시되는 밀도 정보를 합성곱신경망을 이용하는 경우 예측력이 높아졌기에, 향후 이를 개선하는 연구를 기대한다.

House price forecasting is a major topic of interest to a wide range of stakeholders, from individuals to institutions to governments. Until now, predictions have been made using statistical models, but recently, there have been many studies using deep learning and machine learning. In particular, researchers are applying CNN (Convolutional Neural Networks) using image information. However, there have been no studies using image information in Korea. This paper aims to predict apartment prices in Seoul by using Kernel Density Estimation (KDE) to display satellite images and density information of shopping malls and amenities in the area and use them as input variables. In this paper, we collected housing transaction records, local income information, population density, and population data by age group from 2021 to 2022, and collected satellite images divided into zones through Naver Cloud Platform’s Static Map API. In particular, information on shopping malls, hospitals, parks, subway stations, and schools collected from the Open government data portal(www.data.go.kr) and Seoul open data square(data.seoul.go.kr) was converted into density information using KDE. The data collected in various forms was preprocessed to predict the actual transaction price of apartments calculated as a unit price per area. Regression model, multi-layer artificial neural network model, and CNN model were used as prediction models, and their predictive power was compared. The results are as follows. First, among the regression models, the model that adds shopping centre and amenity density information obtained from KDE to demographic and housing-related variables has a high predictive power. Second, the MLP model has a relatively high performance compared to the regression model. Third, the fusion of CNN and MLP performed the best when satellite image features, demographic and housing-related variables, and KDE characteristics were all used, compared to traditional multilayer neural networks and regression models. These results suggest that satellite imagery provides useful information for house price prediction, and we look forward to future research to improve the predictive power of satellite imagery and density information represented by KDE using CNN for house price prediction.

본 논문에서는 음성과 얼굴 정보를 분석하여 자동차환경에서 운전자를 검증하는 기술을 소개한다. 음성정보를 이용한 화자검증을 위해서는 잘 알려진 Mel-scale Frequency Cepstral Coefficients(MFCCs)를 음성 특징으로 사용하였으며, 동영상을 이용한 얼굴검증에 대해서는 AdaBoost를 이용하여 검출된 얼굴 영역에 대해 주성분 분석을 수행하여 데이터의 크기가 현저히 줄어든 특징벡터를 추출하였다. 기존의 화자검증 방식에 비해 본 논문에서는 추출된 음성 및 얼굴 특징들을 Gaussian Mixture Models(GMM)-Supervector기반의 Support Vector Machine(SVM)커넬 방식에 적용하여 운전자의 음성과 얼굴을 효과적으로 검증하는 방식을 제안하였다. 실험결과 제안한 방법은 단순한 GMM 방식이나 SVM 방식보다 운전자 검증성능을 향상시킴을 알 수 있었다.

This paper presents biometrical driver verification system in car experiment through analysis of speech, and face information. We have used Mel-scale Frequency Cesptral Coefficients (MFCCs) for speaker verification using speech information. For face verification, face region is detected by AdaBoost algorithm and dimension-reduced feature vector is extracted by using principal component analysis only from face region. In this paper, we apply the extracted speech- and face feature vectors to an SVM kernel with Gaussian Mixture Models(GMM) supervector. The experimental results of the proposed approach show a clear improvement compared to a simple GMM or SVM approach.

본 연구의 목적은 자기공명 (magnetic resonance, MR) 영상으로부터 각각의 kernel size가 설정된 median modified Wiener filter (MMWF)를 적용하여 그 결과에 대한 노이즈 제거 효과 및 화질 개선 정도를 정량적으로 평가하여 최적의 kernel size를 찾는 것이다. 이를 위해 BrainWeb 시뮬레이션 프로그램을 이용해 노이즈가 부가된 brain T1 강조영상으로부터 kernel size를 3×3, 5×5, 7×7, 9×9, 그리고 11×11으로 변경하며 MMWF를 각각 적용하였다. 정량적 평가를 통한 영상 특성의 변화를 분석하기 위해 contrast to noise ratio (CNR) 및 coefficient of variation (COV)를 측정하였다. 결과적으로, kernel size가 증가함에 따라 CNR은 증가하는 경향을 보이며 11×11 에서 가장 우수한 값을 나타내었고, COV는 9×9에서 가장 개선된 값을 나타내었다. 하지만 육안적 평가 결과, kernel size가 증가함에 따라 조직 간의 경계 구분 능력이 현저히 저하되는 것으로 나타났다. 결론적으로, MR 영상으로부터 MMWF를 적용할 때, 선예도 측면에서는 5×5가, 노이즈 제거 측면에서는 7×7의 kernel size가 가장 적합하다고 판단 된다.

The purpose of this study is to apply a median modified Wiener filter (MMWF) with each kernel size set to a magnetic resonance (MR) image and quantitatively evaluate the noise reduction effect and image quality improvement on the result to optimize the kernel size for MMWF. For this, MMWF was applied by changing the kernel size of 3×3, 5×5, 7×7, 9×9, and 11×11 to the noise-added brain T1-weighted image acquired using the BrainWeb simulation program. To analyze the changes of image characteristics through quantitative evaluation, contrast to noise ratio (CNR) and coefficient of variation (COV) were calculated. As a result of quantitative evaluation, as the kernel size increased, CNR showed a tendency to increase and showed the best value at 11×11, and COV showed the most improved value at 9×9. However, as a result of visual evaluation, it was found that as the kernel size increased, the ability to distinguish boundaries between tissues was significantly reduced. In conclusion, when applying MMWF from MR images, it is judged that a kernel size of 5×5 is most suitable in terms of sharpness and 7×7 in terms of noise removal.

This study examines the empirical performance of three model-based option valuation approaches in the KOSPI200 options market. We evaluate the in-sample pricing, out-of-sample pricing and hedging performance of the approaches based on the specification of option pricing models directly (a pricing model-based approach), on the pricing kernels implied by the option pricing models (an implied pricing kernel-based approach), and on parametric pricing kernels which are independently structured to have their own explicit functional forms (a parametric pricing kernel-based approach). Two option pricing models, a GARCH option pricing model and a Black-Scholes (BS) option pricing model, and their implied pricing kernels are analyzed and two parametric pricing kernel specifications suggested by Rosenberg and Engle (2002) are compared in a unified framework which extends the GARCH process of Duan (1995) to reflect the dynamics of asymmetric volatility. We find that the empirical performance of the approaches related to the GARCH and Black-Scholes option pricing models is moderately improved when we estimate the structural parameters using options data (options-based estimation) compared to the model performances when estimating the parameters using only a time-series of underlying returns data (underlying returns-based estimation). With the estimates under the underlying returns-based estimation, the pricing modelbased option valuation approach outperforms the implied pricing kernel-based option valuation approach for both the GARCH and BS option pricing models. However, with the estimates under the options-based estimation, this relationship is reversed in pricing OTM options in the case of the GARCH option pricing model. Although the BS option pricing model is generally the worst performer with the estimates under the underlying returns-based estimation, it yields better performance for pricing ITM options and similar performance for hedging compared to the GARCH option pricing model with the estimates under the options-based estimation. The option valuation approach based on the parametric pricing kernel of which functional form is a Chebyshev polynomial performs best out of all approaches and methods considered in this study

빗물이나 도로 표면의 종류에 따라 발생되는 반사광은 도로 상황을 자동으로 분석하는 비전 기술에서 가장 어려운 문제점 중의 하나이다. 특히 무인 자동 주행에서는 운전자의 안전을 위하여 어떠한 경우에도 정확한 도로 분석을 필요로 하며, 동시에 실시간 처리도 요구된다. 본 논문에서는 여러 가지 상황에서 조명 반사를 제거하기 위하여 커널 독립 성분 분석을 이용하여 조명 반사와 도로 정보를 분리하는 방법을 제안하며, 실시간 처리를 위하여 CUDA를 이용하여 구현하였다. 제안된 방법은 실험을 통해서 성능 평가 되었으며, 향후 적용 가능한 가능성을 보여주었다.

Lighting reflection caused by rainwater or status of road surface is one of the most difficult problem in automatic vision system. In autonomous vehicle system, accurate analysis on road condition and realtime implementation should be required concurrently. In this paper, we propose the separation of road information and reflection using kernel independent component analysis and its implementation using CUDA for realtime application. The experimental results show that proposed method is plausible and applicable to real application in the future.

Stereo vision is actively researched as a solution for distance measurement in autonomous driving. This technique involves triangulation-based distance calculation using left and right images acquired from two image sensors. A kernel window is employed to determine the disparity between these left and right images. To achieve optimal disparity in various environments, it is essential to support different kernel sizes. Moreover, there is a need for research to integrate high-throughput memory, such as high bandwidth memory (HBM), for processing real-time highresolution stereo vision images from sensors. In this paper, we propose a stereo vision accelerator structure utilizing HBM, which supports various kernel sizes.