• Journal of Intelligence and Information Systems
• /
• v.11 no.1
• /
• pp.169-189
• /
• 2005

In order to derive a better classifier with a limited number of training examples, active teaming alternately repeats the querying stage fur category labeling and the subsequent learning stage fur rebuilding the calssifier with the newly expanded training set. To relieve the user from the burden of labeling, especially in an on-line environment, it is important to minimize the number of querying steps as well as the total number of query examples. We can derive a good classifier in a small number of querying steps by using only a small number of examples if we can select multiple of diverse, representative, and ambiguous examples to present to the user at each querying step. In this paper, we propose a cluster-based batch query selection method which can select diverse, representative, and highly ambiguous examples for efficient active learning. Experiments with various text data sets have shown that our method can derive a better classifier than other methods which only take into account the ambiguity as the criterion to select multiple query examples.

• The Journal of the Korea Contents Association
• /
• v.9 no.10
• /
• pp.13-21
• /
• 2009

This paper describes research of image enhancement for detection of face area. Typical face recognition algorithms used fixed parameter filtering algorithms to optimize face images for the recognition process. A fixed filtering scheme introduces errors when applied to face images captured in various different environmental conditions. For acquiring face image of good quality from the image including complex background and illumination, we propose a method for image enhancement using the categories based on the image intensity values. When an image is acquired average values of image from sub-window are computed and then compared to training values that were computed during preprocessing. The category is selected and the most suitable image filter method is applied to the image. We used histogram equalization, and gamma correction filters with two different parameters, and then used the most suitable filter among those three. An increase in enrollment of filtered images was observed compared to enrollment rates of the original images.

• 한국HCI학회:학술대회논문집
• /
• 2008.02b
• /
• pp.298-303
• /
• 2008

Watching a program on IPTV among the numerous choices from the internet requires a burden of searching and browsing for a favorite one. This paper introduces a new concept called Mosaic Map and presents how it provides preview information of image map links to other programs. In Mosaic Map the pixels in the still image are used both as shading the background and as thumbnails which can link up with other programs. This kind of contextualized preview of choices can help IPTV users to associate the image with related programs without making visual saccades between watching IPTV and browsing many choices. The experiments showed that the Mosaic Map reduces the time to complete search and browsing, comparing to the legacy menu and web search.

• Journal of Korea Society of Industrial Information Systems
• /
• v.17 no.3
• /
• pp.59-68
• /
• 2012

Zero client-based cloud storage technology is gaining much interest as a tool to centralized management of organizational documents nowadays. Besides the well-known cloud storage's defects such as security and privacy protection, users of the zero client-based cloud storage point out the difficulty in browsing and selecting the storage category because of its diversity and complexity. To resolve this problem, this study proposes a method of intelligent document archiving by applying an algorithm-based automatic topic identification technology. Without user's direct definition of category to store the working document, the proposed methodology and prototype enable the working documents to be automatically archived into the predefined categories according to the extracted topic. Based on the proposed ideas, more effective and efficient centralized management of electronic documents can be achieved.

• Korean Journal of Soil Science and Fertilizer
• /
• v.25 no.3
• /
• pp.213-219
• /
• 1992

Geomorphological properties for the phylite-derived soil were examined to relate processes to four landscape positions, shoulder, backslope, footslope, and toeslope in Chungbuk. The distribution of Ogcheon-geology system was 216 thousand ha in South Korea. 2 orders, 3 suborders, 4 great groups, 5 subgroups, and 9 series were mapped. Soil color was interlocked by landscape. Soil color index values and $Fe_2O_3$ contents increased with soil-drainge class. Silt/clay and Ca/Mg ratios tended to narrow wish elevation and decreased with depth. Therefore, profile development or age on the landscape position was shoulder>backslope>footslope>toeslope. Color index(C2m) has a sighificant correlation with $Fe_2O_3$, in soil profile($r=0.777^{**}$). Pedologic type was continuity/discontinuity and soil property changes of depth<12cm would have a continous function.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.2
• /
• pp.166-172
• /
• 2014

Image classification is an important problem in computer vision. However, it is a very challenging problem due to the variability, ambiguity and scale change that exists in images. In this paper, we propose a method of a hypergraph based modeling can consider the higher-order relationships of semantic attributes of a scene image and apply it to a scene image classification. In order to generate the hypergraph optimized for specific scene category, we propose a novel search method based on a probabilistic subspace method and also propose a method to aggregate the expression values of the member semantic attributes that belongs to the searched subsets based on a linear transformation method via likelihood based estimation. To verify the superiority of the proposed method, we showed that the discrimination power of the feature vector generated by the proposed method is better than existing methods through experiments. And also, in a scene classification experiment, the proposed method shows a competitive classification performance compared with the conventional methods.

• The KIPS Transactions:PartD
• /
• v.13D no.7 s.110
• /
• pp.1027-1038
• /
• 2006

There are many technical problems in the recommendation system based on very large database(VLDB). So, it is necessary to study the recommendation system' structure and the data-mining technique suitable for the large scale Internet shopping mail. Thus we design and implement the product recommendation system using k-means clustering algorithm and sequential pattern technique which can be used in large scale Internet shopping mall. This paper processes user information by batch processing, defines the various categories by hierarchical structure, and uses a sequential pattern mining technique for the search engine. For predictive modeling and experiment, we use the real data(user's interest and preference of given category) extracted from log file of the major Internet shopping mall in Korea during 30 days. And we define PRP(Predictive Recommend Precision), PRR(Predictive Recommend Recall), and PF1(Predictive Factor One-measure) for evaluation. In the result of experiments, the best recommendation time and the best learning time of our system are much as O(N) and the values of measures are very excellent.

• Journal of Information Management
• /
• v.40 no.2
• /
• pp.71-94
• /
• 2009

Indexing images using traditional indexing methods like taxonomy is not always efficient because of its visual content. This study examined how to apply folksonomies to image retrieval. To do this, first, we developed a category model for image tags found in Flickr. The model includes five categories and seventeen subcategories. Second, in order to evaluate the usefulness of the model to represent the various image tags as well as to investigate the end-user tagging behavior, three researchers classified the sampled image tags(141 most popular tags, 105 tags on three individual tag clouds and 3,848 image tags assigned on 156 images) according to the model. Finally, based on the research results, we proposed three methods for efficient image retrieval: extending folksonomies by combining them with ontologies; improving image retrieval efficiency using visual content and folksonomies; and updating taxonomy using folksonomies.

• The Journal of Bigdata
• /
• v.6 no.1
• /
• pp.161-178
• /
• 2021

The importance of 'Big Data' is increasing to the point that it is likened to '21st century crude oil'. For smart city IoT data, attention should be paid to quality control as the quality of data is associated with the quality of public services. However, data quality indicators presented through ISO/IEC organizations and domestic/foreign organizations are limited to the 'User' perspective. To complement these limitations, the study derives supplier-centric indicators and their priorities. After deriving 3 categories and 13 indicators of supplier-oriented smart city IoT data quality evaluation indicators, we derived the priority of indicator categories and data quality indicators through AHP analysis and investigated the feasibility of each indicator. The study can contribute to improving sensor data quality by presenting the basic requirements that data should have to individuals or companies performing the task. Furthermore, data quality control can be performed based on indicator priorities to provide improvements in quality control task efficiency.

• Journal of Korean Society of Forest Science
• /
• v.100 no.3
• /
• pp.494-503
• /
• 2011

This study was aimed at applying the IUCN category system to the Baekdudaegan Protected Area. A classification key was developed to apply the system to the overlapped designated protected areas inside of Baekdudaegan Protected Area. Korea national parks and forests managers' and experts' opinions were collected and they all agreed to the use of multiple classification in Baekdudaegan Protected Area. For example, the type of natural forests among the Forest Genetic Resources Reserves was classified to be IUCN Category Ia while other types of Forest Genetic Resources Reserve was classified to be Category IV. And the Protected Forest Landscape was classified to be Category V while the other types of protected forests were classified to be Category VI. The study suggests the need of classification of forest protected areas including Baekdudaegan Protected Area using IUCN system accompanying with protected areas management effectiveness evaluation.