• Proceedings of the Korean Information Science Society Conference
• /
• 1999.10b
• /
• pp.362-364
• /
• 1999

본 논문에서는 Quadtree 구조를 기반으로 한 효율적인 색상 정보 추출과 영상 검색을 수행하는 시스템을 구현한다. 제시한 시스템은 원 영상으로부터 DC영상을 추출하고, DC 영상의 화소들을 RGB에서 HSV의 색상 좌표계로 변환한다. 변환된 영상에서 색상의 분포에 따라 Quadtree 형태로 영역을 분할하고 대표 색상을 추출한다. 마지막으로 추출한 색상과 그 색상의 분포에 따라 Quadtree 형태로 영역을 분할하고 대표 색상을 추출한다. 마지막으로 추출한 색상과 그 색상의 분포값을 영역의 위치에 따라 Quadtree의 단말 노드에 저장한다. 그리고, 사용자가 질의 영상을 주었을 때 Quadtree에 저장된 정보를 이용하여 본 논문에서 제안한 유사도 측정을 통하여 결과 영상을 보여준다. 본 논문에서 제안한 방법으로 실험한 결과 64개의 영역으로 나눈 방법에 비해 비교하는 평균 영역의 개수는 28.9개였고, 검색시간은 2~6초 정도 감소하였다. 또한, 전체 영상의 색상 정보 저장량도 25% 정도 줄어들었지만 질의에 대한 두 방법의 검색 결과는 유사하게 나타났다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.6A
• /
• pp.575-583
• /
• 2002

JBIG2 is an International Standard fur compression of Bi-level images and documents. JBIG2 supports three encoding modes for high compression according to region features of documents. One of which is generic region coding for bitmap coding. The basic bitmap coder is either MMR or arithmetic coding. Pattern matching coding method is used for text region, and halftone pattern coding is used for halftone region. In this paper, a document is segmented into line-art, halftone and text region for JBIG2 encoding and JBIG2 CODEC is implemented. For efficient region segmentation of documents, region segmentation method using wavelet coefficient is applied with existing boundary extraction technique. In case of facsimile test image(IEEE-167a), there is improvement in compression ratio of about 2% and enhancement of subjective quality. Also, we propose arbitrary shape halftone region coding, which improves subjective quality in talc neighboring text of halftone region.

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.39 no.3
• /
• pp.35-41
• /
• 2002

There are many image segmentation methods having been published as the results of research so far, but it is difficult to be partitioned to each similar range that should be extracted into the accurate parameters of image information on the images with noises. Also if it is used to fractal coding, according to amount of noise in image, the image segmentation leads to decreasing of the compression ratio. In this paper, we propose the new quad-tree image segmentation using the box-counting dimension which can estimate the effective image information parameters against the noise properties and apply this method to fractal image coding. As the result of simulation, we confirm that the image segmentation is improved to 31.10% for parameter detection of image information and compression ratio is enhanced to 38.93% for fractal image coding when tested on 10% Gaussian white noise image by the proposed quad-tree method compared with method using existing quad-tree. 

• Journal of Korea Multimedia Society
• /
• v.5 no.6
• /
• pp.616-624
• /
• 2002

An extraction method of central objects in the color images is proposed, in this paper. A central object is defined as a comparatively consist of the central object in the image. First of all. an input image and its decreased resolution images are segmented. Segmented regions are classified as the outer or the inner region. The outer region is adjacent regions are included by a same region in the decreased resolution image. Then core object regions and core background regions are selected from the inner region and the outer region respectively. Core object regions are the representative regions for the object and are selected by using the information about the information about the region size and location. Each inner regions is classified into foreground or background regions by comparing values of a color histogram intersection of the inner region against the core object region and the core background regions. The core object region and foreground regions consist of the central object in the image.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.8
• /
• pp.57-66
• /
• 2011

In this paper, it is proposed for algorithm to detect human efficiently using a edge symmetry and gradient directional characteristics in realtime by the feature extraction in a single input image. Proposed algorithm is composed of three stages, preprocessing, region partition of human candidates, verification of candidate regions. Here, preprocessing stage is strong the image regardless of the intensity and brightness of surrounding environment, also detects a contour with characteristics of human as considering the shape features size and the condition of human for characteristic of human. And stage for region partition of human candidates has separated the region with edge symmetry for human and size in the detected contour, also divided 1st candidates region with applying the adaboost algorithm. Finally, the candidate region verification stage makes excellent the performance for the false detection by verifying the candidate region using feature vector of a gradient for divided local area and classifier. The results of the simulations, which is applying the proposed algorithm, the processing speed of the proposed algorithms is improved approximately 1.7 times, also, the FNR(False Negative Rate) is confirmed to be better 3% than the conventional algorithm which is a single structure algorithm.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.10b
• /
• pp.592-594
• /
• 2003

이 논문에서는 wavelet과 sobel filter를 사용하여 영상의 객관적인 평가 점수를 계산하는 새로운 기법을 제안한다. 이 기법은 orthogonal wavelet 변환을 기초로 하고 있으며 원본 영상과 처리된 영상 데이터가 모두 가용하다는 것을 전제로 한다. Wavelet을 이용해 주파수에 따라 분할된 영상 정보를 이용해 각각의 부영역 별 차영상을 획득하고 이 획득된 영상의 에너지를 이용해 화질 평가 수치를 계산한다. 부영역 별로 획득된 영상은 일정한 크기의 블록으로 분할되어 동일한 블록 내에서 가용한 영상의 특징에 관한 정보(contrast, edge 영역의 분포 정도) 벡터와 내적하여 새로운 특징 벡터로 사용되고, 이 특징 벡터의 가중치를 최적화하여 높은 상관도의 화질평가 점수를 산출하게 된다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2008.11a
• /
• pp.173-176
• /
• 2008

본 논문에서는 한 장의 기준 영상과 그에 상응하는 참 깊이 맵을 이용하여 가상의 다중 시점 영상 생성 시 발생하는 가려짐 영역 보간 방법을 제안한다. 이 방법은 영상 인페인팅 기술과 각각의 깊이 정도에 따른 층별 보간 기술을 이용한다. 우선, 기준영상을 깊이 정보에 따라 여러 개의 층으로 분할한다. 각각의 층에 대해 가려짐 영역 내의 화소들은 영상 인페인팅 기술을 이용하여 보간한다. 마지막 단계 에서 개별적으로 보간 된 층 영상들은 하나로 합성되어 가상 시점의 영상을 이룬다. 영상을 깊이 정보에 따라 분할함으로써, 각 깊이 정도에 대한 텍스쳐의 연관성을 보존하며 보간 할 수 있으므로 기존의 방법에 비하여 보다 정확하고 세밀한 가려짐 영역 보간이 가능하다. 본 논문에서는 여러 가지 실험 결과를 통하여 제안한 방법의 효율성을 입증하였다.

• Journal of Biomedical Engineering Research
• /
• v.18 no.4
• /
• pp.339-348
• /
• 1997

In this paper, we propose a coding method to improve compression efficiency for MR image. This can be achieved by combining coding scheme and segmentation scheme which removes noisy background region, which is meaningless for diagnosis in the MR image. In segmentation algoritm, we use full-resolution wavelet transform to extract features of regions in image and Kohonen self-organizing map to classify the features. The subsequent wavelet coder encodes only diagnostically significant foreground regions refering to segmentation map. Our proposed algorithm provides about 15% of bit rate reduction when compared with the same coder which is not combined with segmentation scheme. And the proposed scheme shows better reconstructed image quality than JPEG at the same compression ratio.

• Journal of Korea Multimedia Society
• /
• v.12 no.11
• /
• pp.1547-1556
• /
• 2009

In this paper, we propose a flame detection method using region expansions and on-line variances in outdoor infrared video sequences. To segment flame candidates' regions in infrared images, we first, extract initial regions by high threshold values in infrared images and then the segmented regions are expanded to their neighbors with similar high intensity values. The segmented regions could be non-flame areas like bare-grounds and buildings. Therefore, to detect flame regions in the segmented regions, the segmented regions which have high intensity values in infrared image, are tracked using bounding regions in frame sequences. Variances in the tracked regions are calculated effectively by on-line updates to measure intensity variations on the tracked regions. Experiments show that the proposed method, which is based on region expansions and the average of on-line variances in the regions, is efficient to detect flames in infrared image.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.10C
• /
• pp.1023-1032
• /
• 2003

Object-based coding, such as MPEG-4, enables various content-based functionalities for multimedia applications. In order to support such functionalities, as well as to improve coding efficiency, each frame of video sequences should be segmented into video objects. In this paper. we propose an effective video object segmentation method using nonlinear multiscale filtering and spatio-temporal information. Proposed method performs a spatial segmentation using a nonlinear multiscale filtering based on the stabilized inverse diffusion equation(SIDE). And, the segmented regions are merged using region adjacency graph(RAG). In this paper, we use a statistical significance test and a time-variant memory as temporal segmentation methods. By combining of extracted spatial and temporal segmentations, we can segment the video objects effectively. Proposed method is more robust to noise than the existing watershed algorithm. Experimental result shows that the proposed method improves a boundary accuracy ratio by 43% on "Akiyo" and by 29% on "Claire" than A. Neri's Method does.