• Journal of KIISE:Software and Applications
• /
• v.26 no.12
• /
• pp.1519-1530
• /
• 1999

오늘날 멀티미디어 정보의 양이 매우 빠른 속도로 증가함에 따라 멀티미디어 데이타베이스에 대한 효율적인 관리는 더욱 중요한 의미를 가지게 되었다. 게다가 영상과 같은 비 문자형태의 데이타에 대한 사용자들의 내용기반 검색욕구 증가로 인해 비디오 인덱싱에 대한 관심은 더욱 고조되고 있다. 따라서 본 논문에서는 우선적으로 분할된 샷 경계면에서 추출된 대표 프레임과 정지 영상 데이타베이스로부터 유사 영상과 유사 대표 프레임을 검색할 수 있는 환경을 제공한다. 우선적으로 영상에 의한 질의는 기존에 주로 사용되어온 색상 히스토그램방식을 탈피하여 본 논문에서 제안하는 CS와 GS방식을 이용하여 색상 및 방향성 정보도 고려하도록 설계하였다. 또한 얼굴에 의한 질의는 대표 프레임으로부터 얼굴 영역을 추출해 내고 얼굴의 경계선 값 및 쌍 직교 웨이블릿 변환에 의해 얻어진 2개의 특징값을 이용하여 유사 인물이 포함된 대표 프레임을 검색해 내도록 설계하였다. Abstract There is a rapid increase in the use of digital video information in recent years, it becomes more important to manage multimedia databases efficiently. There is a big concern about video indexing because users require content-based image retrieval. In this paper, we first propose query-by-image system environment which allows to retrieve similar images from the chosen representative frames or images from the image databases. This algorithm considers not only the discretized color histogram but also the proposed directional information called CS & GS method. Finally, we designe another query environment using query-by-face. In this system , user selects a people in the representative frame browser and then system extracts a face region from that frame. After that system retrieves similar representative frames using 2 features, edge information and biorthogonal wavelet transform.

• Journal of Information Processing Systems
• /
• v.14 no.2
• /
• pp.469-486
• /
• 2018

We propose an enhanced version of the local binary pattern (LBP) operator for texture extraction in images in the context of image retrieval. The novelty of our proposal is based on the observation that the LBP exploits only the lowest kind of local information through the global histogram. However, such global Histograms reflect only the statistical distribution of the various LBP codes in the image. The block based LBP, which uses local histograms of the LBP, was one of few tentative to catch higher level textural information. We believe that important local and useful information in between the two levels is just ignored by the two schemas. The newly developed method: gradual locality integration of binary patterns (GLIBP) is a novel attempt to catch as much local information as possible, in a gradual fashion. Indeed, GLIBP aggregates the texture features present in grayscale images extracted by LBP through a complex structure. The used framework is comprised of a multitude of ellipse-shaped regions that are arranged in circular-concentric forms of increasing size. The framework of ellipses is in fact derived from a simple parameterized generator. In addition, the elliptic forms allow targeting texture directionality, which is a very useful property in texture characterization. In addition, the general framework of ellipses allows for taking into account the spatial information (specifically rotation). The effectiveness of GLIBP was investigated on the Corel-1K (Wang) dataset. It was also compared to published works including the very effective DLEP. Results show significant higher or comparable performance of GLIBP with regard to the other methods, which qualifies it as a good tool for scene images retrieval.

• The Journal of the Korea Contents Association
• /
• v.6 no.10
• /
• pp.134-142
• /
• 2006

In this paper, we investigate a Hausdorff distance, which is used for the measurement of image similarity, to see whether it is also effective for document retrieval. The proposed method uses a local crowdedness and a Hausdorff distance to locate text images by determining whether a pair of images scanned at different time comes from the same text or not. To reduce the processing time, which is one of the disadvantages of a Hausdorff distance algorithm, we adopt a local crowdedness for feature point extraction. We apply the proposed method to 190 pairs of the same class and 190 pairs of the different class collected from postal envelop images. The results show that the modified Hausdorff distance proposed in this paper performed well in locating the tort region and calculating the degree of similarity between two images. An improvement of accuracy by 2.7% and 9.0% has been obtained, compared to a binary correlation method and the original Hausdorff distance method, respectively.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2008.11a
• /
• pp.255-257
• /
• 2008

본 논문에서는 최근 IT 기술의 발전에 따라 무수히 양산되고 있는 멀티미디어 데이터를 효율적으로 검색하기 위한 방법을 제안한다. 영상 검색 시스템에 사용되는 데이터베이스(DB) 영상들에 존재하는 각 객체들의 존재 영역을 기반으로 질의 영상 (query image)의 객체 영역을 추정해서 검색에 활용하는 것이다. 이는 질의 영상의 전체 영역으로부터 객체를 추정하는 것보다 데이터베이스 영상들로부터 추출한 통계적 객체 분포 범위를 기반으로 추정하기 때문에 빨리 객체 추출이 가능하도록 한다. 따라서 객체를 추출하기 위한 배경 지식이나, 사용자 입력이 전혀 필요 없다. 이렇게 추출된 객체 영역의 영상들로부터 GLCM 알고리즘을 이용해서 객체 영역의 특성이 잘 반영된 질감 특징 값을 바탕으로 검색에 활용 할 경우 원본 영상의 질감 특징을 활용한 경우보다, 객체의 질감 특징을 더 잘 반영한다는 것을 실험을 통해 확인할 수 있었다.

• The KIPS Transactions:PartD
• /
• v.15D no.4
• /
• pp.463-474
• /
• 2008

The advancement of information technology and the popularization of Internet has explosively increased the amount of multimedia contents. Therefore, the requirement of multimedia recommendation to satisfy a user's needs increases fastly. Up to now, CF is used to recommend general items and multimedia contents. However, general CF doesn't reflect visual characteristics of image contents so that it can't be adaptable to image recommendation. Besides, it has limitations in new item recommendation, the sparsity problem, and dynamic change of user preference. In this paper, we present new image recommendation method FBCF (Feature Based Collaborative Filtering) to resolve such problems. FBCF builds new user profile by clustering visual features in terms of user preference, and reflects user's current preference to recommendation by using preference feedback. Experimental result using real mobile images demonstrate that FBCF outperforms conventional CF by 400% in terms of recommendation ratio.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.1
• /
• pp.91-97
• /
• 2009

This paper proposes novel FE-CBIRS that finds best position of a cut to be retrieved based on color feature distribution in digital contents of IPTV. Conventional CBIRS have used a method that utilizes both color and shape information together to classify images, as well as a method that utilizes both feature information of the entire region and feature information of a partial region that is extracted by segmentation for searching. Also, in the algorithm, average, standard deviation and skewness values are used in case of color features for each hue, saturation and intensity values respectively. Furthermore, in case of using partial regions, only a few major colors are used and in case of shape features, the invariant moment is mainly used on the extracted partial regions. Due to these reasons, some problems have been issued in CBIRS in processing time and accuracy so far. Therefore, in order to tackle these problems, this paper proposes the FE-CBIRS that makes searching speed faster by classifying and indexing the extracted color information by each class and by using several cuts that are restricted in range as comparative images.

• Korean Journal of Environment and Ecology
• /
• v.22 no.3
• /
• pp.289-300
• /
• 2008

This study aims to discriminate differences in natural landscapes between the Cairngorms National Park in Scotland and the Jirisan National Park in Korea, using functions of content-based image retrieval such as texture, shape, and color. Digital photographs of each National Park were taken and selected. The low-level functions of photographic images were reduced to orthogonally rotated five factors. Based on the reduced factors, a linear decision boundary was obtained between Cairngorms landscapes and Jirisan landscapes. As a result, the discriminant function significantly delineated two groups, resulting in $x^2=63.40$ with df=5(p<0.001). Both the eigenvalue 2.417 and the value of wilks' lambda 0.29 supported that the most proportion of total variability came from the differences between the means of discriminant function of groups. It was estimated that four independent variables explained about 70.7% of total variance of dependent variable. The variable with the largest effect on landscapes was far region-related factor(r=1.07), followed by near region-related factor (r=0.90). A total of 90.7% of cross-validated grouped cases were correctly classified. It was interpreted that far distant regions, as well as near distant regions, had sufficient discrimination power for landscape classification between the Cairngorms National Park and the Jirisan National Park, so that landscape identity of the National Park over cultures was revealed by skylines in a most effective way. Relatively fewer factors making visual landscapes were effectively used to classify natural landscapes of the National Parks which had different semantics.

• Journal of Korea Multimedia Society
• /
• v.4 no.5
• /
• pp.423-429
• /
• 2001

In this paper, we propose the improved method for retrieving images with DC element of DCT that is used in image compression such as JPEG/MPEG. The existing method retrieves images with DC of fixed block size DCT. In this method, the increase in the block size results in faster retrieving speed, but it lessens the accuracy. The decrease in the block size improves the accuracy, however, it degrades the retrieving speed. In order to solve this problem, the proposed method utilizes the variable block size DCT. This method first determines the existence of object regions within each block, and then creates an image region table. Based on this table, it determines the size of each block, following a simple rule; decrease the block size in the object regions, and increase the block size in the background regions. The proposed method using variable block size DCT improves about 15% in terms of the accuracy. Additionally, when there rarely exist images of same pattern, it is able to retrieve faster only by comparing the image region patterns.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2006.05a
• /
• pp.127-130
• /
• 2006

본 논문에서는 MPEG-7 의 특징 기술자를 이용하고, 초기 중요도 가중치를 고려한 관심영역(ROI: Region-Of-Interest) 기반의 의료 영상 검색 시스템을 제안한다. 의료 영상에서 의미 없는 배경 부분을 제거하고, 영역 추출 처리 시간을 줄이는 관심 윈도우(AW: Attention Window)를 생성하여 관심 영역 세그먼테이션을 수행한다. 또한 인간 시각에 부합하는 검색 성능의 향상을 위해 특징 벡터 거리 계산에서 영역의 초기 가중치를 설정하였다. 실험에서 구현된 시스템은 의료 영상을 효과적으로 찾아내며, 조합된 특징과 가중치를 이용한 유사도 측정으로 검색 성능이 향상됨을 보여준다.

• Proceedings of the Korea Society for Industrial Systems Conference
• /
• 1999.05a
• /
• pp.68-74
• /
• 1999

본 논문에서는 영상의 내용을 나타내는 키워드를 이용하는 기존의 텍스트 기반 영상 검색과 영역 색상 정보를 이용한 내용 기반 영상 검색을 결합한 시스템을 구현함으로서, 보다 효과적인 영상 검색을 할 수 있도록 하였다. 영상의 크기는 입력된 원 영상을 사용하였으며, 색상 정보 추출에 있어 HSI 공간으로 변환하여 256개의 칼라로 양자화하였다. 보통의 정지 영상의 경우 대부분의 객체가 중앙에 있을 경우를 고려하여, 영상을 중앙 영역과 배경 영역으로 구분하고, 각각의 영역에서 두 개의 히스토그램을 생성한다. 중앙 영역과 배경영역의 히스토그램 인터섹션을 이용한 검색을 실험하였고, 영역색상과 기존의 키워드를 결합한 검색도 또한 실험하였다. 기존의 히스토그램 인터섹션의 경우 Precision/Recall이 0.34/0.60인데 비해 영역 색상 히스토그램을 인터섹션한 경우의 Precision/Recall은 0.69/0.76이고 키워드를 결합한 경우의 Precision/Recall은 0.92/0.80를 얻음으로써, 제안된 방식의 검색이 비교적 효율적임을 보였다.