• The Korean Journal of Applied Statistics
• /
• v.34 no.4
• /
• pp.599-609
• /
• 2021

Data with a large difference in the number of objects between clusters are called unbalanced data. In discriminant analysis of unbalanced data, it is more important to classify objects in minority categories than to classify objects in majority categories well. However, objects in minority categories are often misclassified into majority categories. In this study, we propose a method that combined hierarchical DBSCAN (HDBSCAN) and SMOTE to solve this problem. Using HDBSCAN, it removes noise in minority categories and majority categories. Then it applies SMOTE to create new data. Area under the roc curve (AUC) and F1 scores were used to compare performance with existing methods. As a result, in most cases, the method combining HDBSCAN and synthetic minority oversampling technique (SMOTE) showed a high performance index, and it was found to be an excellent method for classifying unbalanced data.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.8
• /
• pp.1976-1995
• /
• 2023

The anchor-free object detector CenterNet regards the object as a center point and predicts it based on the Gaussian circle region. For each object's center point, CenterNet directly regresses the width and height of the objects and finally gets the boundary range of the objects. However, the critical range of the object's center point can not be accurately limited by using the Gaussian circle region to constrain the prediction region, resulting in many low-quality centers' predicted values. In addition, because of the large difference between the width and height of different objects, directly regressing the width and height will make the model difficult to converge and lose the intrinsic relationship between them, thereby reducing the stability and consistency of accuracy. For these problems, we proposed a center point prediction method based on the Gaussian elliptic region and a size component regression method based on the small solution space. First, we constructed a Gaussian ellipse region that can accurately predict the object's center point. Second, we recode the width and height of the objects, which significantly reduces the regression solution space and improves the convergence speed of the model. Finally, we jointly decode the predicted components, enhancing the internal relationship between the size components and improving the accuracy consistency. Experiments show that when using CenterNet as the improved baseline and Hourglass-104 as the backbone, on the MS COCO dataset, our improved model achieved 44.7%, which is 2.6% higher than the baseline.

• Journal of KIISE:Software and Applications
• /
• v.37 no.3
• /
• pp.194-204
• /
• 2010

A space, real or virtual, comprises regions as its parts and physical objects residing in them. A coherent and sophisticated representaion scheme for their spatial relations premises the precision and plausibility in its associated agents' inferencing on the spatial relations and the development of events occurring in such a space. The existing spatial models are not suitable for a comprehensive representation of the general spatial relations in that they have limited expressive powers based on the dichotomy between the large and small scales, or support only a small set of topological relations. The representaion model we propose has the following distinctive chracteristics: firstly, our model provides a comprehensive representation scheme to accommodate large and small scale spaces in an integrated fashion; secondly, our model greatly elaborated the spatial relations among the small-scale objects based on their contact relations and the compositional relations among their respective components objects beyond the basic topological relations like disjoint and touch; thirdly, our model further diversifies the types of supported relations by adding the container property besides the soildness together with considering the gravity direction. The resulting integrated spatial knowledge representation scheme considering the gravity allows the diverse spatial relations in the real world to be simulated in a precise manner in relation to the associated spatial events and provides an expression measure for the agents in such a cyber-world to capture the spatial knowledge to be used for recognizing the situations in the spatial aspects.

• Journal of KIISE:Databases
• /
• v.33 no.3
• /
• pp.271-281
• /
• 2006

Recently the need for Location-Based Service (LBS) has increased due to the development and widespread use of the mobile devices (e.g., PDAs, cellular phones, labtop computers, GPS, and RFID etc). The core technology of LBS is a moving-objects database that stores and manages the positions of moving objects. To search for information quickly, the database needs to contain an index that supports both real-time position tracking and management of large numbers of updates. As a result, the index requires a structure operating in the main memory for real-time processing and requires a technique to migrate part of the index from the main memory to disk storage (or from disk storage to the main memory) to manage large volumes of data. To satisfy these requirements, this paper suggests a unified index scheme unifying the main memory and the disk as well as migration policies for migrating part of the index from the memory to the disk during a restriction in memory space. Migration policy determines a group of nodes, called the migration subtree, and migrates the group as a unit to reduce disk I/O. This method takes advantage of bulk operations and dynamic clustering. The unified index is created by applying various migration policies. This paper measures and compares the performance of the migration policies using experimental evaluation.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.29 no.6
• /
• pp.621-629
• /
• 2011

According to the development of technologies for generating the 3D spatial information, the needs for producing and updating the precise 3D objects with LoD 4 level are increased. On the other hand, the needs for real-time updating of 2D digital maps are expanded, based on the execution of various GIS projects. These 2D informations can be extracted from precisely constructed 3D spatial information, to do this the feasibility studies on extraction of the 2D information from the 3D spatial information is needed. In this study, 3D objects are modeled using multi-oblique images, and the objects are stereo-plotted using digital airborne images, as well. Then the two data sets are compared and analyzed. The results show that the accuracy assessments fulfill the 1/1,000 digital map accuracy standard of regulations for photogrametric surveying of National Geographic Information Institute, but the shapes and the areas of building objects are different between two data sets because of the portrayal standards. Consequently, researchers can conclude that it is possible to generate the building layer of large scale topographic map using multi-oblique images, but additional researches is needed to resolve the problems on differences of the portrayal standards.

• Korean Journal of Remote Sensing
• /
• v.37 no.2
• /
• pp.177-198
• /
• 2021

Existing change detection researches have been focused on changes of land use and land cover (LULC), damaged areas, or large vegetated and water regions. On the other hands, increased temporal and spatial resolution of satellite images are strongly suggesting the feasibility of change detection of small objects such as vehicles and ships. In order to check the feasibility, this paper analyzes the performance of existing pixel-based change detection methods over small objects. We applied pixel differencing, PCA (principal component analysis) analysis, MAD (Multivariate Alteration Detection), and IR-MAD (Iteratively Reweighted-MAD) to Kompsat-3A and Google Map images taken within 10 days. We extracted ground references for changed and non-changed small objects from the images and used them for performance analysis of change detection results. Our analysis showed that MAD and IR-MAD, that are known to perform best over LULC and large areal changes, offered best performance over small object changes among the methods tested. It also showed that the spectral band with high reflectivity of the object of interest needs to be included for change analysis.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.9 no.1
• /
• pp.1-12
• /
• 2006

Approaches to spatial analysis differ from the type of spatial objects to be treated. Especially, in here, the case where two spatial data sets coexist is considered. The goal of such case lies on detecting a subset of spatial objects out of a large set that affects the distribution of the other object. However, it is not easy to extract a subset from a large set by visualization just with the help of GIS since huge amount of data are provided nowadays. In this research, therefore, relationship between two different spatial data are analyzed by quantitative measure in the case study of marketing geography. A purchase history data of a small retail store and the location of its competitors are given as source data for the analysis. The goal of analysis from the aspect of this case study is to extract strong competitors of the store that affects the sales amount of the store among many competitors. With the result, therefore, it is expected that market area pattern and competitive power of stores under micro scale retail environment would be understood by quantitative measure.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.13 no.3
• /
• pp.40-47
• /
• 1999

We have established a pulsed low-energy large-aIXTture electron beam(LELAEB) generation system with an energy of 2OO[keV], current of 1[A], pulse repetition rate of 200[Hz], and several tens of ${\mu}s$ pulse width. The system is characterized by a cold cathode that is simpler than the hot cathode. Electron beam does not need to be scanned over target objects because of large beam aIXTture of $300[\textrm{cm}^2]$. Electron source is secondary electrons that are generated when the ions from the glow discharge collide on the cathode surface. In this paper, We report about the design and manufacture of LELAEB generation system based on the theoretical analysis in order to study lXlssibility of increasing the efficiency of IELAEB accelerator. We also report on the possibility of large aperture beam current generation and the current density uniformity based on the experiIrental results.esults.

• Journal of Korea Multimedia Society
• /
• v.7 no.12
• /
• pp.1680-1691
• /
• 2004

One of the representative methods of optical flow is a gradient method which estimates the movement of an object based on the differential of image brightness. However, the method is ineffective for large displacement of the object and many improved methods have been proposed to copy with such limitations. One of these improved techniques is the multigrid processing, which is used in many optical flow algorithms. As an alternative novel technique we have been proposing an orthogonal functional expansion method, where whole displacements are expanded from low frequency terms. This method is expected to be applicable to flow estimation with large displacement and deformation including expansion and contraction, which are difficult to cope with by conventional optical flow methods. In the orthogonal functional expansion method, the apparent displacement field is calculated iteratively by a projection method which utilizes derivatives of the invariant constraint equations of brightness constancy. One feature of this method is that differentiation of the input image is not necessary, thereby reducing sensitivity to noise. In this paper, we apply our method to several real images in which the objects undergo large displacement and/or deformation including expansion. We demonstrate the effectiveness of the orthogonal functional expansion method by comparing with conventional methods including our optimally scaled multigrid optical flow algorithm.

• Smart Structures and Systems
• /
• v.24 no.5
• /
• pp.587-595
• /
• 2019

Deformation measurement of large scale structures, such as the ground beds of high-rise buildings, tunnels, bridge, and railways, are important for insuring service quality and safety. The pose-relay videometrics method and displacement-relay videometrics method have already presented to measure the pose of non-intervisible objects and vertical subsidence of unstable areas, respectively. Both methods combine the cameras and cooperative markers to form the camera series networks. Based on these two networks, we propose two novel videometrics methods with closed-loop camera series network for deformation measurement of large scale structures. The closed-loop camera series network offers "closed-loop constraints" for the camera series network: the deformation of the reference points observed by different measurement stations is identical. The closed-loop constraints improve the measurement accuracy using camera series network. Furthermore, multiple closed-loops and the flexible combination of camera series network are introduced to facilitate more complex deformation measurement tasks. Simulated results show that the closed-loop constraints can enhance the measurement accuracy of camera series network effectively.