• 지능정보연구
• /
• 제10권2호
• /
• pp.15-37
• /
• 2004

최근 데이터 양이 급증하면서 데이터 마이닝에 대한 연구가 활발하게 진행되고 있으며 특히 GPS 시스템, 감시시스템, 기상 관측 시스템과 같은 다양한 응용 시스템으로부터 수집된 데이터를 분석하고자 하는 시공간 데이터 마이닝 연구에 대한 관심이 더욱 높아지고 있다. 기존의 시공간 데이터 마이닝 연구들에서는 비시공간 데이터 기반의 일반적인 클러스터링 기법들을 그대로 적용하고 있으나 데이터의 속성이 다른 시공간 데이터 마이닝에서 기존의 알고리즘들이 어느 정도의 성능을 보장하는지, 데이터의 시공간 속성에 따라 적절한 마이닝 알고리즘을 선택하기 위한 기준이 무엇인지 등에 대한 연구는 미흡한 실정이다. 본 논문에서는 기존의 시공간 데이터 마이닝 연구에서 일반적으로 많이 사용되어 온 알고리즘인 SOM(Self-Organizing Map)을 기반으로 시공간 데이터 마이닝 모듈을 개발하고, 개발된 클러스터링 모듈의 성능을 K-means과 두 가지 응집 계층(Hierarchical Agglomerative) 알고리즘들과 균질도, 분리도, 반면영상 너비, 정확도의 네 가지 평가 기준을 기반으로 비교하였다. 또한 입력 데이터의 특성 가시화 및 클러스터링 결과의 정확한 분석을 위해 시공간 데이터 클러스터링을 위한 가시화 모듈을 개발하였다.

• 한국컴퓨터정보학회논문지
• /
• 제21권9호
• /
• pp.37-43
• /
• 2016

This paper introduces a method for video shot group detection needed for efficient management and summary of video. The proposed method detects shots based on low-level visual properties and performs temporal and spatial clustering based on visual similarity of neighboring shots. Shot groups created from temporal clustering are further clustered into small groups with respect to visual similarity. A set of representative shot frames are selected from each cluster of the smaller groups representing a scene. Shots excluded from temporal clustering are also clustered into groups from which representative shot frames are selected. A number of video clips are collected and applied to the method for accuracy of shot group detection. We achieved 91% of accuracy of the method for shot group detection. The number of representative shot frames is reduced to 1/3 of the total shot frames. The experiment also shows the inverse relationship between accuracy and compression rate.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제3권3호
• /
• pp.331-343
• /
• 2009

Many types of sensor data exhibit strong correlation in both space and time. Both temporal and spatial suppressions provide opportunities for reducing the energy cost of sensor data collection. Unfortunately, existing clustering algorithms are difficult to utilize the spatial or temporal opportunities, because they just organize clusters based on the distribution of sensor nodes or the network topology but not on the correlation of sensor data. In this paper, we propose a novel clustering algorithm based on the correlation of sensor data. We modify the advertisement sub-phase and TDMA schedule scheme to organize clusters by adjacent sensor nodes which have similar readings. Also, we propose a spatio-temporal suppression scheme for our clustering algorithm. In order to show the superiority of our clustering algorithm, we compare it with the existing suppression algorithms in terms of the lifetime of the sensor network and the size of data which have been collected in the base station. As a result, our experimental results show that the size of data is reduced and the whole network lifetime is prolonged.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권7호
• /
• pp.3239-3267
• /
• 2018

Image and video dehazing is a popular topic in the field of computer vision and digital image processing. A fast, optimized dehazing algorithm was recently proposed that enhances contrast and reduces flickering artifacts in a dehazed video sequence by minimizing a cost function that makes transmission values spatially and temporally coherent. However, its fixed-size block partitioning leads to block effects. The temporal cost function also suffers from the temporal non-coherence of newly appearing objects in a scene. Further, the weak edges in a hazy image are not addressed. Hence, a video dehazing algorithm based on well designed spectral clustering is proposed. To avoid block artifacts, the spectral clustering is customized to segment static scenes to ensure the same target has the same transmission value. Assuming that edge images dehazed with optimized transmission values have richer detail than before restoration, an edge intensity function is added to the spatial consistency cost model. Atmospheric light is estimated using a modified quadtree search. Different temporal transmission models are established for newly appearing objects, static backgrounds, and moving objects. The experimental results demonstrate that the new method provides higher dehazing quality and lower time complexity than the previous technique.

• Journal of Electrical Engineering and Technology
• /
• 제13권1호
• /
• pp.38-50
• /
• 2018

Spatio-temporal load forecasting (STLF) is a foundation for building the prediction-based power map, which could be a useful tool for the visualization and tendency assessment of urban energy application. Constructing one point-forecasting model for each electricity cell in the geographic space is possible; however, it is unadvisable and insufficient, considering the aggregation features of electricity cells and uncertainties in input variables. This paper presents a new STLF method, with a data-driven framework consisting of 3 subroutines: multi-level clustering of cells considering their aggregation features, load regression for each category of cells based on SLS-SVRNs (sparse least squares support vector regression networks), and interval forecasting of spatio-temporal load with sampled blind number. Take some area in Pudong, Shanghai as the region of study. Results of multi-level clustering show that electricity cells in the same category are clustered in geographic space to some extent, which reveals the spatial aggregation feature of cells. For cellular load regression, a comparison has been made with 3 other forecasting methods, indicating the higher accuracy of the proposed method in point-forecasting of spatio-temporal load. Furthermore, results of interval load forecasting demonstrate that the proposed prediction-interval construction method can effectively convey the uncertainties in input variables.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권6호
• /
• pp.2217-2229
• /
• 2015

Since the variations of illumination, the irregular changes of human shapes, and the partial occlusions, multiple person tracking is a challenging work in computer vision. In this paper, we propose a graph clustering method based on spatio-temporal information of moving objects for multiple person tracking. First, the part-based model is utilized to localize individual foreground regions in each frame. Then, we heuristically leverage the spatio-temporal constraints to generate a set of reliable tracklets. Finally, the graph shift method is applied to handle tracklet association problem and consequently generate the completed trajectory for individual object. The extensive comparison experiments demonstrate the superiority of the proposed method.

• Journal of information and communication convergence engineering
• /
• 제6권4호
• /
• pp.434-438
• /
• 2008

Massive data are continuously produced with a data rate of over several terabytes every day. These applications need effective clustering algorithms to achieve an overall high performance computation. In this paper, we propose ancestor as cluster center based approach to clustering, the K-means algorithm using ancestor. We modify the K-means algorithm. We present a clustering architecture and a clustering algorithm that minimize of I/Os and show a performance with excellent. In our experimental performance evaluation, we present that our algorithm can improve the I/O speed and the query processing time.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권7호
• /
• pp.2488-2511
• /
• 2015

Data redundancy has high impact on Wireless Sensor Network's (WSN) performance and reliability. Spatial and temporal similarity is an inherent property of sensory data. By reducing this spatio-temporal data redundancy, substantial amount of nodal energy and bandwidth can be conserved. Most of the data gathering approaches use either temporal correlation or spatial correlation to minimize data redundancy. In Collective Prediction exploiting Spatio Temporal correlation (CoPeST), we exploit both the spatial and temporal correlation between sensory data. In the proposed work, the spatial redundancy of sensor data is reduced by similarity based sub clustering, where closely correlated sensor nodes are represented by a single representative node. The temporal redundancy is reduced by model based prediction approach, where only a subset of sensor data is transmitted and the rest is predicted. The proposed work reduces substantial amount of energy expensive communication, while maintaining the data within user define error threshold. Being a distributed approach, the proposed work is highly scalable. The work achieves up to 65% data reduction in a periodical data gathering system with an error tolerance of 0.6℃ on collected data.

• ETRI Journal
• /
• 제37권2호
• /
• pp.317-327
• /
• 2015

A reliable analysis of consumer preference from a large amount of purchase data acquired in real time and an accurate customer characterization technique are essential for successful direct marketing campaigns. In this study, an optimal segmentation of post office customers in Korea is performed using a subspace projection-based clustering method to generate an accurate customer characterization from a high-dimensional census dataset. Moreover, a traditional temporal mining method is extended to an algorithm using the MapReduce framework for a consumer preference analysis. The experimental results show that it is possible to use parallel mining through a MapReduce-based algorithm and that the execution time of the algorithm is faster than that of a traditional method.

• Journal of information and communication convergence engineering
• /
• 제6권2호
• /
• pp.182-186
• /
• 2008

RFID systems generate huge amount of data quickly. The data are associated with the locations and the timestamps and the containment relationships. It is requires to assure efficient queries and updates for product tracking and monitoring. We propose a clustering technique for fast query processing. Our study presents the state charts of temporal event flow and proposes the dependence trees with data association and uses them to cluster the linked events. Our experimental evaluation show the power of proposing clustering technique based on dependence tree.