• Geotechnical Engineering
• /
• v.14 no.2
• /
• pp.107-126
• /
• 1998

Ground settlements induced by tunnel excavation cause the foundations of the neighboring building structures to deform. An expert system called NESASS( Neural network Expert System for Adjacent Structure Safety analysis) was developed to analyze the structural safety of such building structures. NESASS predicts the trend of ground settlements resulting from tunnel excavation and carries out a safety analysis for building structures on the basis of the predicted ground settlements. Using neural network technique. the NESASS learns the database consisting of the measured ground settlements collected from numerous actual fields and infers a settlement trend at the field of interest. The NESASS calculates the magnitudes of angular distortion, deflection ratio, and differential settlement of the structure. and in turn, determines the safety of the structure. In addition, the NESASS predicts the patterns of cracks to be formed in the structure, using Dulacska model for crack evaluation. In this study, the ground settlements measured from Seoul subway construction sites were collected and classified with respect to the major factors influencing ground settlement. Subsequently, a database of ground settlement due to tunnel excavation was built. A parametric study was performed to select the optimal neural network model for the database. A comparison of the ground settlement predicted by the NESASS with the measured ones indicates that the NESASS leads to reasonable predictions. The results of confidence evaluation for safety evaluation system of the NESASS are presented in this paper.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.6
• /
• pp.1087-1098
• /
• 2023

The rapid advancement of artificial intelligence (AI) technology has led to its proactive utilization across various fields. However, this widespread adoption of AI-based systems has raised concerns about the increasing threat of attacks on these systems. In particular, deep neural networks, commonly used in deep learning, have been found vulnerable to adversarial attacks that intentionally manipulate input data to induce model errors. In this study, we propose a method to protect image classification models from visually imperceptible One-Pixel attacks, where only a single pixel is altered in an image. The proposed defense technique utilizes an autoencoder model to remove potential threat elements from input images before forwarding them to the classification model. Experimental results, using the CIFAR-10 dataset, demonstrate that the autoencoder-based defense approach significantly improves the robustness of pretrained image classification models against One-Pixel attacks, with an average defense rate enhancement of 81.2%, all without the need for modifications to the existing models.

• Geophysics and Geophysical Exploration
• /
• v.26 no.4
• /
• pp.171-184
• /
• 2023

Machine learning (ML)-based cavity detection using a large amount of survey data obtained from vehicle-mounted ground penetrating radar (GPR) has been actively studied to identify underground cavities. However, only simple image processing techniques have been used for preprocessing the ML input, and many conventional seismic and GPR data processing techniques, which have been used for decades, have not been fully exploited. In this study, based on the idea that a cavity can be identified using diffraction, we applied ML-based diffraction separation to GPR data to increase the accuracy of cavity detection using the YOLO v5 model. The original ML-based seismic diffraction separation technique was modified, and the separated diffraction image was used as the input to train the cavity detection model. The performance of the proposed method was verified using public GPR data released by the Seoul Metropolitan Government. Underground cavities and objects were more accurately detected using separated diffraction images. In the future, the proposed method can be useful in various fields in which GPR surveys are used.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.12
• /
• pp.519-524
• /
• 2023

In hashing-based image retrieval, the hash code of a manipulated image is different from the original image, making it difficult to search for the same image. This paper proposes and evaluates a self-supervised deephashing model that generates perceptual hash codes from feature information such as texture, shape, and color of images. The comparison models are autoencoder-based variational inference models, but the encoder is designed with a fully connected layer, convolutional neural network, and transformer modules. The proposed model is a variational inference model that includes a SimAM module of extracting geometric patterns and positional relationships within images. The SimAM module can learn latent vectors highlighting objects or local regions through an energy function using the activation values of neurons and surrounding neurons. The proposed method is a representation learning model that can generate low-dimensional latent vectors from high-dimensional input images, and the latent vectors are binarized into distinguishable hash code. From the experimental results on public datasets such as CIFAR-10, ImageNet, and NUS-WIDE, the proposed model is superior to the comparative model and analyzed to have equivalent performance to the supervised learning-based deephashing model. The proposed model can be used in application systems that require low-dimensional representation of images, such as image search or copyright image determination.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.5 s.305
• /
• pp.123-136
• /
• 2005

We propose a recognition system for superimposed patterns based on selective attention model and SVM which produces better performance than artificial neural network. The proposed selective attention model includes attention layer prior to SVM which affects SVM's input parameters. It also behaves as selective filter. The philosophy behind selective attention model is to find the stopping criteria to stop training and also defines the confidence measure of the selective attention's outcome. Support vector represents the other surrounding sample vectors. The support vector closest to the initial input vector in consideration is chosen. Minimal euclidean distance between the modified input vector based on selective attention and the chosen support vector defines the stopping criteria. It is difficult to define the confidence measure of selective attention if we apply common selective attention model, A new way of doffing the confidence measure can be set under the constraint that each modified input pixel does not cross over the boundary of original input pixel, thus the range of applicable information get increased. This method uses the following information; the Euclidean distance between an input pattern and modified pattern, the output of SVM, the support vector output of hidden neuron that is the closest to the initial input pattern. For the recognition experiment, 45 different combinations of USPS digit data are used. Better recognition performance is seen when selective attention is applied along with SVM than SVM only. Also, the proposed selective attention shows better performance than common selective attention.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.38 no.5
• /
• pp.538-547
• /
• 2001

Currently, 2D real-time image coding techniques had great developments and many related products were commercially developed. However, these techniques lack the capability of handling 3D actuality, occurred by the advent of virtual reality, because they handle only the temporal transmission for 2D image. Besides, many 3D virtual reality researches have been studied in computer graphics. Since the graphical researches were limited to the application of artificial models, the 3D actuality for real scene images could not be managed also. Therefore, a new 3D model extraction method based on stereo vision, that can deal with real scene virtual reality, is proposed in this paper. The proposed method adapted a compact genetic algorithm using population-based incremental learning (PBIL) to matching environments, in order to reduce memory consumption and computational time of conventional genetic algorithms. Since the PBIL used a probability vector and competitive learning, the matching algorithm became simple and the computation load was considerably reduced. Moreover, the matching quality was superior than conventional methods. Even if the characteristics of images are changed, stable outputs were obtained without the modification of the matching algorithm.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2006.05a
• /
• pp.473-476
• /
• 2006

인터넷의 발달로 인하여 정보의 양은 시간이 지날수록 폭발적으로 증가하고 있다. 이러한 방대한 정보로부터 정보검색시스템은 사용자에게 너무 많은 검색결과를 제시하여 사용자가 원하는 정보를 찾기 위해 너무 많은 시간을 소요하게 하는 정보의 과적재 문제가 있다. 질의 기반의 문서요약은 정보의 사용자가 원하는 정보의 검색시간을 줄임으로써 정보의 과적재 문제를 해결하는 방법으로서 점차 중요성이 증가하고 있다. 본 논문은 비음수 행렬 인수분해 (NMF, Non-negative Matrix Factorization)과 코사인 유사도를 이용하여 질의 기반의 문서를 요약하는 새로운 방법을 제안하였다. 제안된 방법은 질의와 문서 간에 사전학습이 필요 없다. 또한 문서를 그래프로 변형시키는 복잡한 처리 없이 NMF 에 의해 얻어진 의미 특징(semantic feature)과 의미 변수(semantic variable)로 문서의 고유 구조를 반영하여 요약의 정확도를 높일 수 있다. 마지막으로 단순한 방법으로 문장을 쉽게 요약할 수 있다.

• Journal of Korea Multimedia Society
• /
• v.13 no.11
• /
• pp.1586-1593
• /
• 2010

Vocabulary recognition error correction method has probabilistic pattern matting and dynamic pattern matting. In it's a sentences to based on key-word by semantic analysis. Therefore it has problem with key-word not semantic analysis for morphological changes shape. Recognition rate improve of vocabulary unrecognized reduced this paper is propose. In syllable restoration algorithm find out semantic of a phoneme recognized by a phoneme semantic analysis process. Using to sentences restoration that morphological analysis and morphological analysis. Find out error correction rate using phoneme likelihood and confidence for system parse. When vocabulary recognition perform error correction for error proved vocabulary. system performance comparison as a result of recognition improve represent 2.0% by method using error pattern learning and error pattern matting, vocabulary mean pattern base on method.

• 한국HCI학회:학술대회논문집
• /
• 2007.02a
• /
• pp.920-926
• /
• 2007

새로운 매체와 접촉 시 발생하는 거부감을 최소화 하고 별도의 학습 없이 사용 가능한 직관적 명령 전달 방식의 매개 인터페이스를 제안한다. 제안하는 매개 인터페이스는 3차원 공간에서 사용가능한 가상 마우스와 TV 리모트 컨트롤러의 기능적 결합을 목표로 하고 실버세대들에게 익숙한 매체인 펜을 형태로 삼아 개념적으로 설계되었다. 구체적인 구현은 가속도계의 신호를 분석하거나 펜촉에 레이저 포인터를 추가하여 레이저 포인터의 좌표 변화를 웹캠으로 추적, 인식하는 방법으로 구분하였고 본 논문에서는 가속도계의 경우를 소개한다. 가속도계 신호분석을 통해 마우스의 기능을 모사하고 동작을 감지하는데 발생하는 문제점과 이를 해결하기 위한 기존 연구를 분석하고 동작 중에 중력방향의 수직축이 바뀌면서 발생하는 가속도계 신호의 오류를 보상하기 위해 제안된 Zero Velocity Compensation 방법을 소개한다. ZVC의 결과에 필수적인 저주파의 시계열 신호 실시간 끝점 추출과 동시에 패턴인식을 위한 특징추출 기능을 수행하는 새로운 알고리즘을 제안하며 기존의 방법과 실험적으로 성능을 비교한다. 또한 입력된 가속도계 신호를 학습된 인식기를 통해 인식하는 기존의 연구에서 더 나아가, 마우스의 좌표변화를 짧은 시간동안 가속도 신호의 실시간 분석을 통해 모사하기 위해 변형시킨 알고리즘을 소개한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.10a
• /
• pp.680-682
• /
• 2018

기후변화에 따라 매개 질병의 발병 빈도가 증가하고 있으며 모기와 같은 매개체에 의해 전염되는 매개 질병은 인구집단에 대한 중요한 위협 요인이다. 이러한 질병 관리를 위해 지역별 모기 서식 현황을 모니터링 하는 시스템의 필요성이 강조되고 있다. 하지만 현재의 모기 모니터링은 개체 파악을 위한 분류와 동정을 사람이 직접 수행하기에 오랜 시간이 소요된다. 이 연구는 그러한 문제점을 해결하고 미래 매개곤충 서식 현황 파악 시스템의 기반을 마련하기 위해 심층 신경망(Deep Neural Networks)을 활용하여 한국 주요 매개모기 종 분류를 수행하고 결과를 분석하였다. 종 분류를 위한 모델은 잘 알려진 신경망 모델인 DenseNet(Densely Connected Networks)을 사용하였고 이를 직접 촬영한 모기 데이터와 약간의 변형을 가한 모기 데이터를 사용하여 학습시켰다. 학습 데이터를 각각 5배, 20배, 100배로 증강하여 실제 데이터의 부족을 보완하였으며, 이를 통해 최대 99.48%의 정확도를 달성하였다.