• The KIPS Transactions:PartB
• /
• v.16B no.1
• /
• pp.47-54
• /
• 2009

Automatic object recognition in 3D measuring data is of great interest in many application fields e.g. computer vision, reverse engineering and digital factory. In this paper we present a software tool for a fully automatic object detection and parameter estimation in unordered and noisy point clouds with a large number of data points. The software consists of three interactive modules each for model selection, point segmentation and model fitting, in which the orthogonal distance fitting (ODF) plays an important role. The ODF algorithms estimate model parameters by minimizing the square sum of the shortest distances between model feature and measurement points. The local quadric surface fitted through ODF to a randomly touched small initial patch of the point cloud provides the necessary initial information for the overall procedures of model selection, point segmentation and model fitting. The performance of the presented software tool will be demonstrated by applying to point clouds.

• Proceedings of the KIEE Conference
• /
• 2002.07a
• /
• pp.320-326
• /
• 2002

Distribution system loss minimization re-configuration is 0-1 planning problem, and the number of combinations requiring searches is extremely large when dealing with typical system scales. For this reason, the application of a genetic algorithm (GA) seems attactive to solve this problem. Although Genetic algorithms are a type of random number search method, they incorporate a multi-point search feature and are therefore superior to one-point search techniques. The efficiency of GAs for solving large combinational problem has received wide attention. Further, parallel searching can be performed and the optimal solution is more easily reached. In this paper, for improving GA convergence characteristics in the distribution system loss minimization re-configeration problem, a chromosome "Limited Life" concept is intro duced. Briefly, considering the population homogenization and genetic drift problems, natural selection is achieved by providing this new concept, in addition to natural selection by fitness. This is possible because individuals in a population have an age value. Simulations were carried out using a model system to check this method's validity.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.360-362
• /
• 2003

This paper presents simple feature-based approaches for full- and/or semi-automatic extraction, selection, and localization (center-determination) of ground control points (GCPs) for radargrammetry using airborne synthetic aperture radar (SAR) images. Test results using airborne NASA/JPL TOPSAR images in Taiwan verify that the registration accuracy is about 0.8${\sim}$1.4 pixels. In c.a. 30 minutes, 1500${\sim}$3000 GCPs are extracted and their point centers in a SAR image of about 512 ${\times}$ 512 pixels are determined on a personal computer.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.6
• /
• pp.1828-1847
• /
• 2014

Coordinated multi-point (CoMP) transmission has been regarded as a potential technology for LTE-Advanced. In frequency division duplexing systems, channel quantization is applied for reporting channel state information (CSI). Considering the dynamic number of cooperation base stations (BSs), asymmetry feature of CoMP channels and high searching complexity, simply increasing the size of the codebook used in traditional multiple antenna systems to quantize the global CSI of CoMP systems directly is infeasible. Per-cell codebook based channel quantization to quantize local CSI for each BS separately is an effective method. In this paper, the theoretical upper bounds of system throughput are derived for two codeword selection schemes, independent codeword selection (ICS) and joint codeword selection (JCS), respectively. The feedback overhead and selection complexity of these two schemes are analyzed. In the simulation, the system throughput of ICS and JCS is compared. Both analysis and simulation results show that JCS has a better tradeoff between system throughput and feedback overhead. The ICS has obvious advantage in complexity, but it needs additional phase information (PI) feedback for obtaining the approximate system throughput with JCS. Under the same number of feedback bits constraint, allocating the number of bits for channel direction information (CDI) and PI quantization can increase the system throughput, but ICS is still inferior to JCS. Based on theoretical analysis and simulation results, some recommendations are given with regard to the application of each scheme respectively.

• Journal of Korea Multimedia Society
• /
• v.25 no.2
• /
• pp.176-188
• /
• 2022

In this paper, we propose a compact method for anthropomorphism that uses Deep Convolutional Neural Networks (DCNN) to detect the similarities between a human face and an animal face. We also apply texture feature-based morphing between them. We propose a basic texture feature-based morphing system for morphing between human faces only. The entire anthropomorphism process starts with the creation of an animal face classifier using a parallel DCNN that determines the most similar animal face to a given human face. The significance of our network is that it contains four sets of convolutional functions that run in parallel, allowing it to extract more features than a linear DCNN network. Our employed texture feature algorithm-based automatic morphing system recognizes the facial features of the human face and takes the Control Points automatically, rather than the traditional human aiding manual morphing system, once the similarity was established. The simulation results show that our suggested DCNN surpasses its competitors with a 92.0% accuracy rate. It also ensures that the most similar animal classes are found, and the texture-based morphing technology automatically completes the morphing process, ensuring a smooth transition from one image to another.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.11
• /
• pp.2849-2865
• /
• 2012

In order to block mobile junk messages automatically, many studies on spam filters have applied machine learning algorithms. Most previous research focused only on the accuracy rate of spam filters from the view point of the algorithm used, not on individual user's preferences. In terms of individual taste, the spam filters implemented on a mobile device have the advantage over spam filters on a network node, because it deals with only incoming messages on the users' phone and generates no additional traffic during the filtering process. However, a spam filter on a mobile phone has to consider the consumption of resources, because energy, memory and computing ability are limited. Moreover, as time passes an increasing number of feature words are likely to exhaust mobile resources. In this paper we propose a spam filter model distributed between a users' computer and smart phone. We expect the model to follow personal decision boundaries and use the uniform resources of smart phones. An authorized user's computer takes on the more complex and time consuming jobs, such as feature selection and training, while the smart phone performs only the minimum amount of work for filtering and utilizes the results of the information calculated on the desktop. Our experiments show that the accuracy of our method is more than 95% with Na$\ddot{i}$ve Bayes and Support Vector Machine, and our model that uses uniform memory does not affect other applications that run on the smart phone.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.9
• /
• pp.4429-4447
• /
• 2018

Local binary descriptors are well-suited for many real-time and/or large-scale computer vision applications, while their low computational complexity is usually accompanied by the limitation of performance. In this paper, we propose a new optimization framework, RLDB (Robust-LDB), to improve a typical region-based binary descriptor LDB (local difference binary) and maintain its computational simplicity. RLDB extends the multi-feature strategy of LDB and applies a more complete region-comparing configuration. A cascade bit selection method is utilized to select the more representative patterns from massive comparison pairs and an online learning strategy further optimizes descriptor for each specific patch separately. They both incorporate LDP (linear discriminant projections) principle to jointly guarantee the robustness and distinctiveness of the features from various scales. Experimental results demonstrate that this integrated learning framework significantly enhances LDB. The improved descriptor achieves a performance comparable to floating-point descriptors on many benchmarks and retains a high computing speed similar to most binary descriptors, which better satisfies the demands of applications.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.4
• /
• pp.563-568
• /
• 2007

This paper presents a methodology for predicting nonlinear time series based on the neural network with weighted fuzzy membership functions (NEWFM). The degree of classification intensity is obtained by bounded sum of weighted fuzzy membership functions extracted by NEWFM, then weighted average defuzzification is used for predicting nonlinear time series. The experimental results demonstrate that NEWFM has the classification capability of 92.22% against the target class of GDP. The time series created by NEWFM model has a relatively close approximation to the GDP which is a typical business cycle indicator, and has been proved to be a useful indicator which has the turning point forecasting capability of average 12 months in the peak point and average 6 months in the trough point during 5th to 8th cyclical period. In addition, NEWFM measures the efficiency of the economic indexes by the feature selection and enables the users to forecast with reduced numbers of 7 among 10 leading indexes while improving the classification rate from 90% to 92.22%.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.4
• /
• pp.1562-1578
• /
• 2020

Vector data compression algorithm can meet requirements of different levels and scales by reducing the data amount of vector graphics, so as to reduce the transmission, processing time and storage overhead of data. In view of the fact that large threshold leading to comparatively large error in Douglas-Peucker vector data compression algorithm, which has difficulty in maintaining the uncertainty of shape features and threshold selection, a segmented Douglas-Peucker algorithm based on node importance is proposed. Firstly, the algorithm uses the vertical chord ratio as the main feature to detect and extract the critical points with large contribution to the shape of the curve, so as to ensure its basic shape. Then, combined with the radial distance constraint, it selects the maximum point as the critical point, and introduces the threshold related to the scale to merge and adjust the critical points, so as to realize local feature extraction between two critical points to meet the requirements in accuracy. Finally, through a large number of different vector data sets, the improved algorithm is analyzed and evaluated from qualitative and quantitative aspects. Experimental results indicate that the improved vector data compression algorithm is better than Douglas-Peucker algorithm in shape retention, compression error, results simplification and time efficiency.

• The KIPS Transactions:PartB
• /
• v.13B no.5 s.108
• /
• pp.499-506
• /
• 2006

This Paper describes a fast and stable camera pose estimation method for real-time augmented reality systems. From the feature tracking results of a marker on a single frame, we estimate the camera rotation matrix and the translation vector. For the camera pose estimation, we use the shape factorization method based on the scaled orthographic Projection model. In the scaled orthographic factorization method, all feature points of an object are assumed roughly at the same distance from the camera, which means the selected reference point and the object shape affect the accuracy of the estimation. This paper proposes a flexible and stable selection method for the reference point. Based on the proposed method, we implemented a video augmentation system that inserts virtual 3D objects into the input video frames. Experimental results showed that the proposed camera pose estimation method is fast and robust relative to the previous methods and it is applicable to various augmented reality applications.