• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.1855-1858
• /
• 2003

In this paper, we propose image retrieval method more accurate and efficient than the conventional one. First of ail, we perform a shot detection and key frame extraction from the DC image constructed by DCT DC coefficients in the compressed video stream that is video compression standard such as MPEG[I][2]. We get principal axis applying PCA(Principal Component Analysis) to key frames for obtaining indexing information, and divide a domain. Video retrieval uses indexing information of high dimension. We apply KD-Trees(K Dimensional-Trees)[3] which shows efficient retrieval in data set of high dimension to video retrieval method. The proposed method can represent property of images more efficiently and property of domains more accurately using KD-Trees.

• Journal of the Korean Society for information Management
• /
• v.28 no.2
• /
• pp.135-158
• /
• 2011

The ability of the searcher to express their information problem to an information retrieval system is fundamental to the retrieval process. Query by visual example is the principal query paradigm for expressing queries in a content-based image retrieval environment yet there is little empirical evidence to support its efficacy in facilitating query formulation. The aim of this research was to investigate the usability of the query by colour method in supporting a range of information problems in order to contribute to the gap in knowledge regarding the relationship between searchers' information problems and the query methods required to support efficient and effective visual query formulation. The results strongly suggest that the query method does not support visual query formulation and that there is a significant mismatch between the searchers information problems and the expressive power of the retrieval paradigm.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.4
• /
• pp.720-732
• /
• 1996

Spatial match retrieval methods for iconic image databases recognize an image document as several icon symbols. Therefore the iconic symbols are used as primary keys to index the image document. When a user requires content-based retrieval ofimages, a spatial match retrieval method converts a query image into iconic symbols and then retrieves relevant images by accessing stored images. In order to support content-based image retrieval efficiently, we, in this paper, propose spatial match retrieval methods using signatures for iconic image databases. For this, we design new index representations of two-dimensional iconic images and explain implemented system.. In addition, we compare the conventional 9-DLT and our two-dimensional image retrieval method in terms of retrieval precision and recall ratio. We show that our method is more efficient than the conventional method.

• Journal of the Semiconductor & Display Technology
• /
• v.7 no.4
• /
• pp.57-62
• /
• 2008

Image retrieval is one of the most exciting and rapidly growing research issues in the field of multimedia technology. This paper proposes a new method that performs search the relevant images by using query-by-example. The proposed method for search and retrieval of images utilizes the location information where the image had been taken. The system associates the photo images with their corresponding GPS coordinates that are used as metadata for searching. Experimental results show that the proposed method demonstrates better performance improving up to 59% of average recall and 49% of average precision. Moreover, we learned from the experimental results geo-location information embedded within the image header is more effective and positive on the search and storage.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.6
• /
• pp.65-71
• /
• 2009

The basic features in the indexing and retrieval of the image is used color, shape, and texture in traditional image retrieval method. We do not use these features and offers a new way. For content-based video indexing and retrieval, visual features used to measure the similarity of the geometric method is presented. This method is called the Radon transform. Without separation, this method is calculated based on the geometric distribution of image. In the experiment has a very good search results.

• Journal of the Korean Society for Library and Information Science
• /
• v.21
• /
• pp.249-290
• /
• 1991

Theoretical developments in the information retrieval have offered a number of alternatives to traditional Boolean retrieval. Probability theory and fuzzy set theory have played prominent roles here. Fuzzy set theory is an attempt to generalize traditional set theory by permitting partial membership in a set and this means recognizing different degrees to which a document can match a request. In this study, an experimentation of a document retrieval system using the fuzzy relation matrix of the keywords is described and the results are offered. The queries composed of keywords and Boolean operaters AND, OR, NOT were processed in the retrieval method, and the method was implemented on the PC of 32bit level (30 MHz) in an experimental system. The measurement of the recall ratio and precision ratio verified the effectiveness of the proposed fuzzy relation matrix of keywords and retrieval method. Compared to traditional crisp method in the same document database, the recall ratio increased $10\%$ high although the precision ratio decreased slightly. The problems, in this experiment, to be resolved are first, the design of the automatic data input and fuzzy indexing modules, through which the system . can have the ability of competition and usefulness. Second, devising a systematic procedure for assigning fuzzy weights to keywords in documents and in queries.

• Journal of the Institute of Convergence Signal Processing
• /
• v.2 no.1
• /
• pp.28-37
• /
• 2001

An image retrieval method combining shape information of 2-dimension color histograms with color information of HSI color histograms is proposed in this paper. In addition, the proposed method can find location information of image through the comparison of similarity among subimages. The suggested retrieval method applies the location information to shape and color information and can retrieve region information which is hard to distinguish in the binary image. Some simulation results show that it works very well in the behalf of precision/recall graph compare with conventional method which uses color histogram. Especially, the proposed method brought well effects such as rotations and transitions of the objects in an image was found.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.3
• /
• pp.522-538
• /
• 2013

Content based image retrieval has become an increasingly important research topic for its wide application. It is highly challenging when facing to large-scale database with large variance. The retrieval systems rely on a key component, the predefined or learned similarity measures over images. We note that, the similarity measures can be potential improved if the data distribution information is exploited using a more sophisticated way. In this paper, we propose a similarity measure learning approach for image retrieval. The similarity measure, so called Fisher kernel, is derived from the probabilistic distribution of images and is the function over observed data, hidden variable and model parameters, where the hidden variables encode high level information which are powerful in discrimination and are failed to be exploited in previous methods. We further propose a discriminative learning method for the similarity measure, i.e., encouraging the learned similarity to take a large value for a pair of images with the same label and to take a small value for a pair of images with distinct labels. The learned similarity measure, fully exploiting the data distribution, is well adapted to dataset and would improve the retrieval system. We evaluate the proposed method on Corel-1000, Corel5k, Caltech101 and MIRFlickr 25,000 databases. The results show the competitive performance of the proposed method.

• The KIPS Transactions:PartD
• /
• v.8D no.5
• /
• pp.451-460
• /
• 2001

In this paper, we propose an indexing model for efficient structure retrieval of XML documents. The proposed indexing model consists of structured information that supports a wide range of queries such as content-based queries and structure-attribute queries at all levels of the document hierarchy and index organizations that are constructed based on the information. To support structured retrieval, a new representation method for structured information is presented. Using this structured information, we design content index, structure index, and attribute index for efficient retrieval. also, we explain processing procedures for mixed queries and evaluate the performance of proposed indexing model. It is shown that the proposed indexing model achieves better retrieval performance than the existing method.

• Journal of the Korean Mathematical Society
• /
• v.36 no.3
• /
• pp.609-619
• /
• 1999

We give a mathematically rigorous description of the recently popular latent semantic indexing (LSI) method in text information retrieval theory. Also, a related problem of finding a document ranking function in linear relevance feedback is discussed.