• The Korean Journal of Applied Statistics
• /
• v.22 no.2
• /
• pp.415-424
• /
• 2009

As one of the unsupervised learning neural network methods, self-organizing maps(SOM) are applied to various fields. It reduces the dimension of multidimensional data by representing observations on the low dimensional manifold. On the other hand, the minimal spanning tree(MST) of a graph that achieves the most economic subset of edges connecting all components by a single open loop. In this study, we apply the MST technique to SOM with subnodes. We propose SOM's with embedded MST and a distance measure for optimum choice of the size and shape of the map. We demonstrate the method with Fisher's Iris data and a real gene expression data. Simulated data sets are also analyzed to check the validity of the proposed method.

• KIPS Transactions on Computer and Communication Systems
• /
• v.9 no.10
• /
• pp.231-238
• /
• 2020

With a large number of documents appearing on the web, document-level novelty detection has become important since it can reduce the efforts of finding novel documents by discarding documents sharing redundant information already seen. A recent work proposed a convolutional neural network (CNN)-based novelty detection model with significant performance improvements. We observed that it has a restriction of using document-level information in determining novelty but assumed that the document-level information is more important. As a solution, this paper proposed two methods of effectively incorporating document-level information using a CNN-based novelty detection model. Our methods focus on constructing a feature vector of a target document to be classified by extracting relative information between the target document and source documents given as evidence. A series of experiments showed the superiority of our methods on a standard benchmark collection, TAP-DLND 1.0.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.1
• /
• pp.45-53
• /
• 2020

Conventional Convolutional Neural Network(CNN) algorithms have limitations in detecting small objects in large image. In this paper, we propose an improved model which is based on Region Of Interest(ROI) selection and image dividing technique. We prepared YOLOv3 / Faster R-CNN algorithms which are transfer-learned by airfield and aircraft datasets. Also we prepared large images for testing. In order to verify our model, we selected airfield area from large image as ROI first and divided it in two power n orders. Then we compared the aircraft detection rates by number of divisions. We could get the best size of divided image pieces for efficient small object detection derived from the comparison of aircraft detection rates. As a result, we could verify that the improved CNN algorithm can detect small object in large images.

• Journal of the Korea Society for Simulation
• /
• v.28 no.4
• /
• pp.21-32
• /
• 2019

Inspired by the social behavior models of a bird flock or fish school, particle swarm optimization (PSO) is a popular metaheuristic optimization algorithm and has been widely used from solving a complex optimization problem to learning a artificial neural network. However, PSO is difficult to apply to many real-life optimization problems involving stochastic noise, since it is originated in a deterministic environment. To resolve this problem, this paper incorporates a resampling method called the uncertainty evaluation (UE) method into PSO. The UE method allows the particles to converge on the accurate optimal solution quickly in a noisy environment by selecting the particles' global best position correctly, one of the significant factors in the performance of PSO. The results of comparative experiments on several benchmark problems demonstrated the improved performance of the propose algorithm compared to the existing studies. In addition, the results of the case study emphasize the necessity of this work. The proposed algorithm is expected to be effectively applied to optimize complex systems through digital twins in the fourth industrial revolution.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.6
• /
• pp.792-798
• /
• 2007

In this paper, we propose a hierarchical fuzzy-rough classification method based on statistical information for maximizing the performance of pattern classification and reducing the number of rules without learning approaches such as neural network, genetic algorithm. In the proposed method, statistical information is used for extracting the partition intervals of antecedent fuzzy sets at each layer on hierarchical fuzzy-rough classification systems and rough sets are used for minimizing the number of fuzzy if-then rules which are associated with the partition intervals extracted by statistical information. To show the effectiveness of the proposed method, we compared the classification results(e.g. the classification accuracy and the number of rules) of the proposed with those of the conventional methods on the Fisher's IRIS data. From the experimental results, we can confirm the fact that the proposed method considers only statistical information of the given data is similar to the classification performance of the conventional methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.2
• /
• pp.800-816
• /
• 2018

In recent years, video surveillance research has been able to recognize various behaviors of pedestrians and analyze the overall situation of objects by combining image analysis technology and deep learning method. Human Activity Recognition (HAR), which is important issue in video surveillance research, is a field to detect abnormal behavior of pedestrians in CCTV environment. In order to recognize human behavior, it is necessary to detect the human in the image and to estimate the pose from the detected human. In this paper, we propose a novel approach for 2D Human Pose Estimation based on object detection using RGB-D information. By adding depth information to the RGB information that has some limitation in detecting object due to lack of topological information, we can improve the detecting accuracy. Subsequently, the rescaled region of the detected object is applied to ConVol.utional Pose Machines (CPM) which is a sequential prediction structure based on ConVol.utional Neural Network. We utilize CPM to generate belief maps to predict the positions of keypoint representing human body parts and to estimate human pose by detecting 14 key body points. From the experimental results, we can prove that the proposed method detects target objects robustly in occlusion. It is also possible to perform 2D human pose estimation by providing an accurately detected region as an input of the CPM. As for the future work, we will estimate the 3D human pose by mapping the 2D coordinate information on the body part onto the 3D space. Consequently, we can provide useful human behavior information in the research of HAR.

• Journal of Korea Multimedia Society
• /
• v.6 no.1
• /
• pp.136-147
• /
• 2003

To achieve an optimal performance for a given problem, we need an integrative process of the parameter optimization via learning and the structure optimization via model selection. In this paper, we propose an efficient optimization method for improving generalization performance by considering the property of each sub-method and by combining them with common theoretical properties. First, weight parameters are optimized by natural gradient teaming with adaptive regularization, which uses a diverse error function. Second, the network structure is optimized by eliminating unnecessary parameters with natural pruning. Through iterating these processes, candidate models are constructed and evaluated based on the Bayesian Information Criterion so that an optimal one is finally selected. Through computational experiments on benchmark problems, we confirm the weight parameter and structure optimization performance of the proposed method.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.4
• /
• pp.669-677
• /
• 2020

As the number of Internet users exploded, attacks on the web increased. In addition, the attack patterns have been diversified to bypass existing defense techniques. Traditional web firewalls are difficult to detect attacks of unknown patterns.Therefore, the method of detecting abnormal behavior by artificial intelligence has been studied as an alternative. Specifically, attempts have been made to apply natural language processing techniques because the type of script or query being exploited consists of text. However, because there are many unknown words in scripts and queries, natural language processing requires a different approach. In this paper, we propose a new classification model which uses byte pair encoding (BPE) technology to learn the embedding vector, that is often used for web attack payloads, and uses an attention mechanism-based Bi-GRU neural network to extract a set of tokens that learn their order and importance. For major web attacks such as SQL injection, cross-site scripting, and command injection attacks, the accuracy of the proposed classification method is about 0.9990 and its accuracy outperforms the model suggested in the previous study.

• The Journal of the Korea Contents Association
• /
• v.20 no.7
• /
• pp.86-96
• /
• 2020

Based on the use cases of image recognition technology, this study looked at how artificial intelligence plays a role in image recognition technology. Through image recognition technology, satellite images can be analyzed with artificial intelligence to reveal the calculation of oil storage tanks in certain countries. And image recognition technology makes it possible for searching images or products similar to images taken or downloaded by users, as well as arranging fruit yields, or detecting plant diseases. Based on deep learning and neural network algorithms, we can recognize people's age, gender, and mood, confirming that image recognition technology is being applied in various industries. In this study, we can look at the use cases of domestic and overseas image recognition technology, as well as see which methods are being applied to the industry. In addition, through this study, the direction of future research was presented, focusing on various successful cases in which image recognition technology was implemented and applied in various industries. At the conclusion, it can be considered that the direction in which domestic image recognition technology should move forward in the future.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.4
• /
• pp.34-40
• /
• 2018

This paper presents a new method to estimate the pose of a moving object accurately using a monocular camera and a low-end GPS+IMU sensor system. For this goal, we adopted a deep neural network for the semantic segmentation of input images and compared the results with a semantic map of a neighborhood. In this map matching, we use weight tables to deal with label inconsistency effectively. Signals from a low-end GPS+IMU sensor system are used to limit search spaces and minimize the proposed function. For the evaluation, we added noise to the signals from a high-end GPS-IMU system. The results show that the pose can be recovered from the noisy signals. We also show that the proposed method is effective in handling non-open-sky situations.