• 한국농촌건축학회논문집
• /
• 제13권1호
• /
• pp.21-28
• /
• 2011

The main purpose of this study is to analyze construction of space through the investigation of the cases of school community library through Gangneung-si's life-learning project and the findings from the analysis could be summarized as below. Firstly, most space used for school community library has the size of two classes in school on average and locals use generally space for reference and learning at school community library. Secondly, the construction of space of school community library is categorized into one for book-returning, references, reading, group learning and information, and an audio-visual space is also used for group learning and reading. A space for book-returning has features based on the location of its entrance and a space for reading features stand-up and sitting-on space considering size and usability. And a space for group learning has the feature of space planning that makes it possible for local people to get library programs and seminars and a space for information shows its feature of space planning that uses the wall.

• 한국인터넷방송통신학회논문지
• /
• 제18권5호
• /
• pp.171-177
• /
• 2018

최근 많은 이미지 데이터 셋들은 일반적인 특성을 추출하기 위한 다양한 데이터 클래스와 특징을 가지고 있다. 하지만 이러한 다양한 데이터 클래스와 특징으로 인해 해당 데이터 셋으로 훈련된 물체 검출 딥러닝 모델은 데이터 특성이 다른 환경에서 좋은 성능을 내지 못하는 단점을 보인다. 이 논문에서는 하위 카테고리 기반 물체 검출 방법과 오픈셋 물체 검출 방법을 이용하여 이를 극복하고, 강인한 물체 검출 딥러닝 모델을 훈련하기 위해 능동 준지도 학습 (Active Semi-Supervised Learning)을 이용한 다중 분기 트리 구조를 제안한다. 우리는 이 구조를 이용함으로써 데이터 특성이 다른 환경에서 적응할 수 있는 모델을 가질 수 있고, 나아가 이 모델을 이용하여 이전의 모델보다 높은 성능을 확보 할 수 있다.

• ETRI Journal
• /
• 제42권1호
• /
• pp.101-107
• /
• 2020

For Korean language processing, morphological analysis is a critical component that requires extensive work. This morphological analysis can be conducted in an end-to-end manner without requiring a complicated feature design using a sequence-to-sequence model. However, the sequence-to-sequence model has a time complexity of O(n²) for an input length n when using the attention mechanism technique for high performance. In this study, we propose a linear-time Korean morphological analysis model using a local monotonic attention mechanism relying on monotonic alignment, which is a characteristic of Korean morphological analysis. The proposed model indicates an extreme improvement in a single threaded environment and a high morphometric F1-measure even for a hard attention model with the elimination of the attention mechanism formula.

• 한국멀티미디어학회논문지
• /
• 제19권9호
• /
• pp.1659-1668
• /
• 2016

Understanding 3D structure of scenes is of a great interest in various vision-related tasks. In this paper, we present a unified approach for estimating depth from a single monocular image. The key idea of our approach is to take advantages both of parametric learning and non-parametric sampling method. Using a parametric convolutional network, our approach learns the relation of various monocular cues, which make a coarse global prediction. We also leverage the local prediction to refine the global prediction. It is practically estimated in a non-parametric framework. The integration of local and global predictions is accomplished by concatenating the feature maps of the global prediction with those from local ones. Experimental results demonstrate that the proposed method outperforms state-of-the-art methods both qualitatively and quantitatively.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권2호
• /
• pp.786-799
• /
• 2021

Road damage detection is important for road maintenance. With the development of deep learning, more and more road damage detection methods have been proposed, such as Fast R-CNN, Faster R-CNN, Mask R-CNN and RetinaNet. However, because shallow and deep layers cannot be extracted at the same time, the existing methods do not perform well in detecting objects with fewer samples. In addition, these methods cannot obtain a highly accurate detecting bounding box. This paper presents a Multi-level Feature Pyramids method based on M2det. Because the feature layer has multi-scale and multi-level architecture, the feature layer containing more information and obvious features can be extracted. Moreover, an attention mechanism is used to improve the accuracy of local boundary boxes in the dataset. Experimental results show that the proposed method is better than the current state-of-the-art methods.

• 한국정보처리학회논문지
• /
• 제6권5호
• /
• pp.1393-1402
• /
• 1999

본 논문에서는 새로운 학습알고리즘의 3층 전향 신경망을 이용한 입력데이터의 주요 특징추출에 대해서 제안하였다. 제안된 학습알고리즘에서에서는 빠른 수렴속도의 최적화가 가능하도록 하기 위하여 기울기하강의 역전파 알고리즘을 이용하고, 국소최적해를 만났을 때 이를 벗어난 새로운 연결가중치의 설정을 위하여 동적터널링의 역전파 알고리즘을 이용함으로써 빠른 수렴속도로 전역최적해로에 수렴되도록 학습시킬 수 있다. 제안된 학습 알고리즘을 이용한 다층신경망을 $12{\times}12$ 픽셀의 영상 데이터들과 $128{\times}128$ 픽셀의 Lenna 영상데이터를 대상으로 시뮬레이션한 결과, 단층신경망을 이용하는 Sanger 방법이나 측면연결을 가지는 단충신경망을 이용하는 Foldiak 방법 및 기울기하강에 기초를 둔 기존의 역전파 알고리즘을 이용한 다층신경망에 의한 결과와 비교할 때 더욱 우수한 수렴성능과 추출성능이 있음을 확인할 수 있었다.

• KEPCO Journal on Electric Power and Energy
• /
• 제2권2호
• /
• pp.311-316
• /
• 2016

인공신경망과 같은 기계학습에 기반한 네트워크 침입탐지/방지시스템은 특징 조합에 따라 탐지의 정확성과 효율성 측면에서 크게 영향을 받는다. 하지만 침입탐지에 사용 가능한 여러개의 특징들 중 정확성과 효율성 측면에서 최적의 특징 조합을 추출하는 특징 선택 문제는 많은 계산량을 요구한다. 본 논문에서는 NSL-KDD 데이터 집합에서 제공하는 6가지 서비스 거부 공격과 정상 트래픽을 구분해 내기 위한 최적 특징 조합 선택 문제를 다룬다. 최적 특징 조합 선택 문제를 해결하기 위해 대표적인 메타 휴리스틱 알고리즘 중 하나인 다중 시작 지역탐색 알고리즘에 기반한 최적 특징 선택 알고리즘을 제시한다. 제안한 특징 선택 알고리즘의 성능 평가를 위해 NSL-KDD 데이터를 상대로 41개의 특징 모두를 사용한 경우와 비교한다. 그리고 선택된 특징 조합을 사용했을 때 가장 높은 성능을 보여주는 기계학습 방법을 찾기위해 3가지 잘 알려진 기계학습 방법들 (베이즈 분류기와 인공신경망, 서포트 벡터 머신)을 사용해 성능을 비교한다.

• 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.

• Journal of Information Processing Systems
• /
• 제11권4호
• /
• pp.643-654
• /
• 2015

Face recognition under controlled settings, such as limited viewpoint and illumination change, can achieve good performance nowadays. However, real world application for face recognition is still challenging. In this paper, we propose using the combination of Affine Scale Invariant Feature Transform (SIFT) and Probabilistic Similarity for face recognition under a large viewpoint change. Affine SIFT is an extension of SIFT algorithm to detect affine invariant local descriptors. Affine SIFT generates a series of different viewpoints using affine transformation. In this way, it allows for a viewpoint difference between the gallery face and probe face. However, the human face is not planar as it contains significant 3D depth. Affine SIFT does not work well for significant change in pose. To complement this, we combined it with probabilistic similarity, which gets the log likelihood between the probe and gallery face based on sum of squared difference (SSD) distribution in an offline learning process. Our experiment results show that our framework achieves impressive better recognition accuracy than other algorithms compared on the FERET database.

• 한국멀티미디어학회논문지
• /
• 제25권1호
• /
• pp.72-81
• /
• 2022

Visual Object Tracking is known as the most fundamental problem in the field of computer vision. Object tracking localize the region of target object with bounding box in the video. In this paper, a custom CNN is created to extract object feature that has strong and various information. This network was constructed as a Siamese network for use as a feature extractor. The input images are passed convolution block composed of a bottleneck layers, and features are emphasized. The feature map of the target object and the search area, extracted from the Siamese network, was input as a local proposal network. Estimate the object area using the feature map. The performance of the tracking algorithm was evaluated using the OTB2013 dataset. Success Plot and Precision Plot were used as evaluation matrix. As a result of the experiment, 0.611 in Success Plot and 0.831 in Precision Plot were achieved.