• The Journal of Korea Robotics Society
• /
• v.4 no.3
• /
• pp.185-191
• /
• 2009

This paper presents a new shape-based algorithm using affine category shape model for object category recognition and model learning. Affine category shape model is a graph of interconnected nodes whose geometric interactions are modeled using pairwise potentials. In its learning phase, it can efficiently handle large pose variations of objects in training images by estimating 2-D homography transformation between the model and the training images. Since the pairwise potentials are defined on only relative geometric relationship betweenfeatures, the proposed matching algorithm is translation and in-plane rotation invariant and robust to affine transformation. We apply spectral matching algorithm to find feature correspondences, which are then used as initial correspondences for RANSAC algorithm. The 2-D homography transformation and the inlier correspondences which are consistent with this estimate can be efficiently estimated through RANSAC, and new correspondences also can be detected by using the estimated 2-D homography transformation. Experimental results on object category database show that the proposed algorithm is robust to pose variation of objects and provides good recognition performance.

• Computers and Concrete
• /
• v.33 no.3
• /
• pp.285-299
• /
• 2024

Recently, many engineering computations have realized their digital transformation to Machine Learning (ML)-based systems. Predicting the behavior of a structure, which is mainly computed with structural analysis software, is an essential step before construction for efficient structural analysis. Especially in the seismic-based design procedure of the structures, predicting the lateral load capacity of reinforced concrete (RC) columns is a vital factor. In this study, a novel ML-based model is proposed to predict the maximum lateral load capacity of RC columns under varying axial loads or cyclic loadings. The proposed model is generated with a Deep Neural Network (DNN) and compared with traditional ML techniques as well as a popular commercial structural analysis software. In the design and test phases of the proposed model, 319 columns with rectangular and square cross-sections are incorporated. In this study, 33 parameters are used to predict the maximum lateral load capacity of each RC column. While some traditional ML techniques perform better prediction than the compared commercial software, the proposed DNN model provides the best prediction results within the analysis. The experimental results reveal the fact that the performance of the proposed DNN model can definitely be used for other engineering purposes as well.

• The Journal of Korea Robotics Society
• /
• v.15 no.1
• /
• pp.1-7
• /
• 2020

In this paper, we introduce the methodology that utilizes deep learning-based front-end to enhance underwater feature matching. Both optical camera and sonar are widely applicable sensors in underwater research, however, each sensor has its own weaknesses, such as light condition and turbidity for the optic camera, and noise for sonar. To overcome the problems, we proposed the opti-acoustic transformation method. Since feature detection in sonar image is challenging, we converted the sonar image to an optic style image. Maintaining the main contents in the sonar image, CNN-based style transfer method changed the style of the image that facilitates feature detection. Finally, we verified our result using cosine similarity comparison and feature matching against the original optic image.

• Journal of Multimedia Information System
• /
• v.6 no.4
• /
• pp.235-238
• /
• 2019

In real world object detection is an active research topic for understanding different objects from images. There are different models presented in past and had significant results. In this paper we are presenting vehicle logo detection using previous object detection models such as You only look once (YOLO) and Faster Region-based CNN (F-RCNN). Both the front and rear view of the vehicles were used for training and testing the proposed method. Along with deep learning an image pre-processing algorithm called perspective transformation is proposed for all the test images. Using perspective transformation, the top view images were transformed into front view images. This algorithm has higher detection rate as compared to raw images. Furthermore, YOLO model has better result as compare to F-RCNN model.

• 한국정보교육학회:학술대회논문집
• /
• 2021.08a
• /
• pp.361-368
• /
• 2021

Daily life such as society, economy, and culture is fundamentally changing due to COVID-19, and digital transformation based on information technology (IT) such as artificial intelligence, data, and cloud is accelerating. In this study, we focused on the metaverse, which is based on the interaction between the virtual world and the real world, and explored the possibility of using the metaverse-based platform for education. The metaverse-based platform was approached from the perspective of the online education ecosystem, which means that not only online teaching and learning activities but also holistic educational activities such as learning, communication, and empathy are performed within the metaverse. In this metaverse platform, learners can feel the presence of learning, and learning motivation and immersion can be promoted. In addition, it is possible to experience self-directed learning based on the autonomy of spatial movement. Although there are technical and ethical limitations to applying the metaverse platform, it would be preferable to focus more on the interaction between learners in the metaverse world rather than high expectations.

• Annual Conference on Human and Language Technology
• /
• 2009.10a
• /
• pp.31-34
• /
• 2009

본 연구의 목표는 비교 문장들을 일곱 가지 유형으로 자동 분류하는 것으로서, 비교 문장 추출, 비교 문장 유형 분류, 유형별 비교 관계 분석으로 이어지는 비교마이닝 세 단계 중 두 번째 과제이다. 본 연구에서는 변환 기반 학습(Transformation-based Learning) 기법을 이용한다. 자연어 처리 분야 여러 부문에서 사용되고 있는 변환 기반 학습은 오류를 감소시키는 최적의 규칙을 자동으로 생성하여 정답을 찾는 규칙 기반 학습 방법이다. 웹상의 다양한 도메인에서 추출한 비교 문장들을 대상으로 실험한 결과, 일곱 가지 비교 문장 유형을 분류하는데 있어서 정확도 80.01%의 우수한 성능을 산출하였다.

• Journal of The Korean Association of Information Education
• /
• v.20 no.3
• /
• pp.313-322
• /
• 2016

The effect on creative problem solving ability and learning flow is analyzed by applying game-based learning using sandbox game, Minecraft Edu for elementary school students. It appeared to be effective when applied to sand box utilizing game-based learning than traditional lecture teaching method on creative problem solving ability and learning flow. It is found to be a significant difference observed in all sub-elements on Creative problem solving ability and it is found to be a significant difference in all sub-elements on learning flow except sense of control and transformation of time.

• Journal of KIISE:Software and Applications
• /
• v.37 no.9
• /
• pp.706-711
• /
• 2010

In the applications of a human speech interface, reducing the error rate in recognition is the one of the main research issues. Many previous studies attempted to correct errors using post-processing, which is dependent on a manually constructed corpus and correction patterns. We propose an automatically learnable post-processing method that is independent of the characteristics of both the domain and the speech recognizer. We divide the entire post-processing task into two steps: error detection and error correction. We consider the error detection step as a classification problem for which we apply the conditional random fields (CRFs) classifier. Furthermore, we apply transformation-based learning (TBL) to the error correction step. Our experimental results indicate that the proposed method corrects a speech recognizer's insertion, deletion, and substitution errors by 25.85%, 3.57%, and 7.42%, respectively.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.2-4
• /
• 2021

In a distributed machine learning system, sharing gradients was considered safe because it did not share original training data. However, recent studies found that malicious attacker could completely restore the original training data from shared gradients. Gradient Leakage Attack is a technique that restoring original training data by exploiting theses vulnerability. In this study, we present the image transformation method based on Discrete Cosine Transform to defend against the Gradient Leakage Attack on the federated learning setting, which training in local devices and sharing gradients to the server. Experiment shows that our image transformation method cannot be completely restored the original data from Gradient Leakage Attack.

• The Journal of the Korea Contents Association
• /
• v.10 no.10
• /
• pp.479-493
• /
• 2010

Along with the transformation of the knowledge-based environment, e-learning has become a main teaching and learning method, prompting various research efforts to be conducted in this field. One major research area in e-learning involves adaptive learning systems that provide personalized learning content according to each learner's characteristics by taking into consideration a variety of learning circumstances. Active research on ontology-based adaptive learning systems has recently been conducted to provide more efficient and adaptive learning content. In this paper, we design and propose an adaptive learning system based on the concept lattice of Formal Concept Analysis (FCA) with the same objectives as those of ontology approaches. However, we are in pursuit of a system that is suitable for learning of specific domains and one that allows users to more freely and easily build their own adaptive learning systems. The proposed system automatically classifies the learning objects and concepts of an evolved domain in the structure of a concept lattice based on the relationships between the objects and concepts. In addition, the system adaptively constructs and presents the learning structure of the concept lattice according to each student's level of knowledge, learning style, learning preference and the learning state of each concept.