• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권2호
• /
• pp.775-792
• /
• 2015

Person re-identification is an important and challenging task in computer vision with numerous real world applications. Despite significant progress has been made in the past few years, person re-identification remains an unsolved problem. This paper presents a novel appearance-based approach to person re-identification. The approach exploits region covariance matrix and color histograms to capture the statistical properties and chromatic information of each object. Robustness against low resolution, viewpoint changes and pose variations is achieved by a novel signature, that is, the combination of Log Covariance Matrix feature and HSV histogram (LCMH). In order to further improve re-identification performance, third-party image sets are utilized as a common reference to sufficiently represent any image set with the same type. Distinctive and reliable features for a given image set are extracted through decision boundary between the specific set and a third-party image set supervised by max-margin criteria. This method enables the usage of an existing dataset to represent new image data without time-consuming data collection and annotation. Comparisons with state-of-the-art methods carried out on benchmark datasets demonstrate promising performance of our method.

• International journal of advanced smart convergence
• /
• 제9권2호
• /
• pp.20-27
• /
• 2020

Sound event detection in real-world environments suffers from the interference of non-stationary and time-varying noise. This paper presents an adaptive noise reduction method for sound event detection based on non-negative matrix factorization (NMF). In this paper, we proposed a deep learning model that integrates Convolution Neural Network (CNN) with Non-Negative Matrix Factorization (NMF). To improve the separation quality of the NMF, it includes noise update technique that learns and adapts the characteristics of the current noise in real time. The noise update technique analyzes the sparsity and activity of the noise bias at the present time and decides the update training based on the noise candidate group obtained every frame in the previous noise reduction stage. Noise bias ranks selected as candidates for update training are updated in real time with discrimination NMF training. This NMF was applied to CNN and Hidden Markov Model(HMM) to achieve improvement for performance of sound event detection. Since CNN has a more obvious performance improvement effect, it can be widely used in sound source based CNN algorithm.

• ETRI Journal
• /
• 제40권4호
• /
• pp.499-510
• /
• 2018

A vector space based augmented structural kinematic (VSASK) feature descriptor is proposed for human activity recognition. An action descriptor is built by integrating the structural and kinematic properties of the actor using vector space based augmented matrix representation. Using the local or global information separately may not provide sufficient action characteristics. The proposed action descriptor combines both the local (pose) and global (position and velocity) features using augmented matrix schema and thereby increases the robustness of the descriptor. A multiclass support vector machine (SVM) is used to learn each action descriptor for the corresponding activity classification and understanding. The performance of the proposed descriptor is experimentally analyzed using the Weizmann and KTH datasets. The average recognition rate for the Weizmann and KTH datasets is 100% and 99.89%, respectively. The computational time for the proposed descriptor learning is 0.003 seconds, which is an improvement of approximately 1.4% over the existing methods.

• 한국정밀공학회지
• /
• 제20권2호
• /
• pp.112-119
• /
• 2003

Generally, PID controller is not suitable to control multi variable system because it is very difficult to tune the PID gains. However, this paper shows that it is not hard to tune the PID gains if we can find a Jacobian matrix of the system. The Jacobian matrix expresses the ratio of output variations according to input variations. It is possible to adjust the input values in order to reduce the output error using the Jacobian. When the colt function is composed of error related terms, the gradient approach can tune the PID gains to minimize the function. In simulation, a hydrofoil catamaran with two inputs and two outputs is applied as a multi variable system. We can easily get the multi variable PID controller by the proposed teaming method. When the controller is compared with LQR controller, the performance is as good as that of LQR controller with a modeling equation.

• 한국HCI학회:학술대회논문집
• /
• 한국HCI학회 2007년도 학술대회 3부
• /
• pp.322-325
• /
• 2007

Graphics processing unit (GPU) is used for a faster artificial neural network. It is used to implement the matrix multiplication of a neural network to enhance the time performance of a text detection system. Preliminary results produced a 20-fold performance enhancement using an ATI RADEON 9700 PRO board. The parallelism of a GPU is fully utilized by accumulating a lot of input feature vectors and weight vectors, then converting the many inner-product operations into one matrix operation. Further research areas include benchmarking the performance with various hardware and GPU-aware learning algorithms. (c) 2004 Pattern Recognition Society. Published by Elsevier Ltd. All rights reserved.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2022년도 추계학술대회
• /
• pp.123-125
• /
• 2022

본 논문에서는 Yolo-pose를 이용하여 장단기 메모리(Long short-term Memory)에 적용하는 시스템을 소개한다. 영상데이터로부터 Yolo-pose를 이용하여 일상생활과 낙상으로 구분된 데이터를 추출하여 LSTM에 적용하여 학습시킨다. 학습은 오버피팅을 방지하기 위하여 8대2의 Validation을 진행하며 Confusion matrix로 나타낸다. Yolo-pose의 결과값은 sensitivity와 specificity 모두 100%를 기록하여 일상생활과 낙상을 잘 구분하는 것을 확인 하였다.

• 대한원격탐사학회지
• /
• 제35권6_1호
• /
• pp.999-1009
• /
• 2019

북극항로의 개척 가능성과 정확한 기후 예측 모델의 필요성에 의해 북극해 고해상도 해빙 지도의 중요성이 증가하고 있다. 그러나 기존의 북극 해빙 지도는 제작에 사용된 위성 영상 취득 센서의 특성에 따른 데이터의 취득과 공간해상도 등에서 그 활용도가 제한된다. 본 연구에서는 Sentinel-1 A/B SAR 위성자료로부터 고해상도 해빙 지도를 생성하기 위한 딥러닝 기반의 해빙 분류 알고리즘을 연구하였다. 북극해 Ice Chart를 기반으로 전문가 판독에 의해 Open Water, First Year Ice, Multi Year Ice의 세 클래스로 구성된 훈련자료를 구축하였으며, Convolutional Neural Network 기반의 두 가지 딥러닝 모델(Simple CNN, Resnet50)과 입사각 및 thermal noise가 보정된 HV 밴드를 포함하는 다섯 가지 입력 밴드 조합을 이용하여 총 10가지 케이스의 해빙 분류를 실시하였다. 이 케이스들에 대하여 Ground Truth Point를 사용하여 정확도를 비교하고, 가장 높은 정확도가 나온 케이스에 대해 confusion matrix 및 Cohen의 kappa 분석을 실시하였다. 또한 전통적으로 분류를 위해 많이 활용되어 온 Maximum Likelihood Classifier 기법을 이용한 분류결과에 대해서도 같은 비교를 하였다. 그 결과 Convolution 층 2개, Max Pooling 층 2개를 가진 구조의 Convolutional Neural Network에 [HV, 입사각] 밴드를 넣은 딥러닝 알고리즘의 분류 결과가 96.66%의 가장 높은 분류 정확도를 보였으며, Cohen의 kappa 계수는 0.9499로 나타나 딥러닝에 의한 해빙 분류는 비교적 높은 분류 결과를 보였다. 또한 모든 딥러닝 케이스는 Maximum Likelihood Classifier 기법에 비해 높은 분류 정확도를 보였다.

• 융합정보논문지
• /
• 제11권11호
• /
• pp.335-344
• /
• 2021

본 논문의 목적은 교과기반 학습성취평가 시스템 구현을 위한 교과 데이터 모델 및 학습 성취도 산출 알고리즘 개발이다. 현재 대학 교육의 방향인 역량기반 교육을 위해서는 교과기반 학습성취 평가가 필수적이지만 기존 연구들은 교육학적 관점으로서 컴퓨터 시스템 관점의 해결책이 매우 부족하다. 본 논문에서는 코스맵 데이터 구조 분석을 통해 계층 구조의 학습성과 모델, 학습모듈 및 학습활동 모델, 학습성과와 학습활동 연계 매트릭스 모델 및 자동화된 성취도 산출 및 성취수준 평가를 위한 성취도 계산 알고리즘을 제안한다. 이를 통해 교과기반 학습성취 평가 시스템을 개발할 수 있으며 시스템 활용을 통해 학습자의 역량 성취를 효과적으로 평가할 수 있다. 제안된 모델과 알고리즘의 평가를 위해 실제 운영중인 자바프로그래밍 교과목에 적용하였으며 이를 통해 교과기반 학습성과 성취평가 시스템 구현의 핵심요소로 활용할 수 있음을 확인하였다. 향후 연구는 학습성과 성취도 산출을 기반으로 적응형 학습 피드백과 개인화된 학습 추천 알고리즘 개발 및 시스템 구현이다.

• 대한방사선기술학회지:방사선기술과학
• /
• 제45권2호
• /
• pp.111-116
• /
• 2022

In this study, the applicability of machine learning for the development of a simulator for general X-ray examination education is evaluated. To this end, k-nearest neighbor(kNN), support vector machine(SVM) and neural network(NN) classification models are analyzed to present the most suitable model by analyzing the results. Image data was obtained by taking 100 photos each corresponding to Posterior anterior(PA), Posterior anterior oblique(Obl), Lateral(Lat), Fan lateral(Fan lat). 70% of the acquired 400 image data were used as training sets for learning machine learning models and 30% were used as test sets for evaluation. and prediction model was constructed for right-handed PA, Obl, Lat, Fan lat image classification. Based on the data set, after constructing the classification model using the kNN, SVM, and NN models, each model was compared through an error matrix. As a result of the evaluation, the accuracy of kNN was 0.967 area under curve(AUC) was 0.993, and the accuracy of SVM was 0.992 AUC was 1.000. The accuracy of NN was 0.992 and AUC was 0.999, which was slightly lower in kNN, but all three models recorded high accuracy and AUC. In this study, right-handed PA, Obl, Lat, Fan lat images were classified and predicted using the machine learning classification models, kNN, SVM, and NN models. The prediction showed that SVM and NN were the same at 0.992, and AUC was similar at 1.000 and 0.999, indicating that both models showed high predictive power and were applicable to educational simulators.

• Computers and Concrete
• /
• 제32권3호
• /
• pp.327-337
• /
• 2023

Concrete is the most widely used building material, with various types including high- and ultra-high-strength, reinforced, normal, and lightweight concretes. However, accurately predicting concrete properties is challenging due to the geotechnical design code's requirement for specific characteristics. To overcome this issue, researchers have turned to new technologies like machine learning to develop proper methodologies for concrete specification. In this study, we propose a highly accurate deep learning-based predictive model to investigate the compressive strength (UCS) of lightweight concrete with natural aggregates (pumice). Our model was implemented on a database containing 249 experimental records and revealed that water, cement, water-cement ratio, fine-coarse aggregate, aggregate substitution rate, fine aggregate replacement, and superplasticizer are the most influential covariates on UCS. To validate our model, we trained and tested it on random subsets of the database, and its performance was evaluated using a confusion matrix and receiver operating characteristic (ROC) overall accuracy. The proposed model was compared with widely known machine learning methods such as MLP, SVM, and DT classifiers to assess its capability. In addition, the model was tested on 25 laboratory UCS tests to evaluate its predictability. Our findings showed that the proposed model achieved the highest accuracy (accuracy=0.97, precision=0.97) and the lowest error rate with a high learning rate (R2=0.914), as confirmed by ROC (AUC=0.971), which is higher than other classifiers. Therefore, the proposed method demonstrates a high level of performance and capability for UCS predictions.