• The Journal of the Korea Contents Association
• /
• v.4 no.1
• /
• pp.18-23
• /
• 2004

In this paper, we propose a new visual feature extraction method for content-based image retrieval(CBIR) based on wavelet transform which has both spatial-frequency characteristic and multi-resolution characteristic. We extract visual features for each frequency band in wavelet transformation and use them to CBIR. The lowest frequency band involves spacial information of original image. We extract L feature vectors using fuzzy homogeneity in the wavelet domain, which consider both the wavelet coefficients and the spacial information of each coefficient. Also, we extract 3 feature vectors wing the energy values of high frequency bands, and store those to image database. As a query, we retrieve the most similar image from image database according to the 10 largest homograms(normalized fuzzy homogeneity vectors) and 3 energy values. Simulation results show that the proposed method has good accuracy in image retrieval using 90 texture images.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.2C
• /
• pp.141-146
• /
• 2003

In this paper, we propose an adaptive search range decision algorithm for motion vector estimation in video coding. The performance of general motion estimation method in video coding mechanism is evaluated with respect to the motion vector accuracy and the complexity, which is trade-off. The proposed algorithm that plays as a role of pre-processing for motion vector estimation determines the motion search range by the local statistics of motion vector of neighboring blocks, resulting in more than 60(%) reduction of the computational cost without the loss of visual quality. Experimental results show the capability of the proposed algorithm.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.5
• /
• pp.56-69
• /
• 1999

본 논문에서는 움직임 보상 차영상 ( motion compensated difference image; MCDI)을 대역 분할한 뒤, 움직임 정보를 이용하여 에너지가 큰 부분만을 선택하여 적응적으로 벡터 양자화를 행함으로써 효율적으로 동영상을 부호화 할 수 있는 기법을 제안하였다. 제안한 기법에서는 MCDI의 특성들을 고려하여 에너지가 낮은 부분은 부호화를 행하지 않고, 에너지가 높은 부분에 대하여서만 움직임 벡터의 크기에 따라 클래스를 분류한 뒤 각 클래스에 맞는 코드북을 설계하여 벡터 양자화를 행한다. 이를 위해서는 부호화 되는 에너지가 높은 부분에 대한 정보와 클래스 부분에 대한 정보가 필요하다. 본 논문에서는 이러한 부가 정보의 양을 움직임 벡터를 이용하여 최소화함으로써 비트율을 크게 줄이면서도 우수한 화질의 복원 영상을 얻을 수 있었다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.7 no.4
• /
• pp.18-28
• /
• 1997

This paper presents image data compression using a classified vector quantization (CVQ) which categories edge blocks according to the energy distribution of subimages in the discrete cosine transform domain. Classifying the edge blocks enhances visual quality of the compressed images while maintaining a high compression ratio. The proposed classification method categories subimages into eight lypes of edge features according to an energy distribution. A neural network, trained with the data generated from the proposed classification method, can successfully classify subimages to eight edge categories. Experimental results are given to show how the (1VQ method incorporatd with a neural network can produce faithful compressed image quality for high compression ratios.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.7
• /
• pp.1518-1523
• /
• 2012

Due to development of visual media in various industrial sectors, the importance of image processing is increasing. Among the various image processing areas, edge detection is utilized widely for various fields such as object recognition, object segmentation, the medical and other industries. Edge includes the critical factors of images like size, direction and location. Then conventional methods such as Sobel, Prewitt, Roberts and Laplacian are proposed to detect edge. However, edge detection property of these methods is declined when they are applied to the image which corrupted by AWGN(Additive White Gaussian Noise). Therefore, an algorithm using modified weighted filter is proposed in this paper and our method has excellent property on edge detection.

• Proceedings of the KAIS Fall Conference
• /
• 2000.10a
• /
• pp.208-213
• /
• 2000

본 논문은 정지 화상에 대한 CBIR(Content-Based Image Retrieval)방법 중 칼라 특성을 이용해서 영상 내 공간 정보를 충분하게 표현할 수 있는 알고리즘을 제안한다. 일반적으로 칼라 특성을 이용한 CBIR은 영상 내 공간정보를 충분하게 표현하지 못하는 단점을 지니고 있다. 이에 기존 논문에서는 인위적으로 영상을 여러 개로 분할하는 방법 등으로 공간정보를 표현하고자 하였지만 특징벡터의 수가 급격히 늘어남에 따라 검색효율이 저하된다는 단점을 가지고있다. 본 논문에서는 기존의 방법을 칼라 오브젝트의 추출 방법에 따라 1차와 2차 관계에 의한 방법으로 분류하고, 이동, 회전 특히 크기 변화(축소, 확대)에 탁월한 성능을 보이는 칼라 오브젝트의 3차 관계를 이용한 방법을 소개한다. 주어진 영상으로부터 양자화된 24개의 버킷을 생성해서 각 버킷 내의 칼라에 대한 색의 표준 편차로 색의 분산 정도틀 나타내고, 히스토그램의 빈도수가 높은 세 개 버킷의 평균 칼라 위치를 계산해서 그들의 상호 각도를 추출하여 영상의 특징 벡터로 사용한을 제안하였다. 실험결과 기존 방법보다 특히 영상의 크기 변화에 대해 좋은 결과를 얻을 수 있었으며, 계산량도 적어 효율적임을 보여 주었다.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2000.08a
• /
• pp.241-244
• /
• 2000

본 논문에서는 차량 영상의 수평 및 수직 명암 값 변화 정보를 이용하여 번호판 영역을 추출하고 원형 패턴 벡터를 이용하여 번호판 내용을 인식하는 알고리즘에 관해 기술하였다. 제안된 알고리즘에서는 번호판 영역에서 문자와 배경이 뚜렷하게 구별되고, 일정한 명암도 변화를 가지면서 다른 영역보다 밀집도가 높다는 특성을 이용하여 수평 및 수직 명암도 변화값을 구하여 차량영상에서 번호판 영역을 추출하며 상당히 어둡거나 밝게 입력된 영상에도 동일한 인식 성능을 얻기 위하여 밝기 보정을 수행한다. 또한, 입력 문자의 크기, 이동 및 회전에 무관한 특성을 추출을 위해 원형 패턴 벡터를 이용하여 차량 번호를 인식하는 알고리즘을 제안하였다. 제안한 방법들을 적용한 결과 계산 속도가 훨씬 빠르며, 차량 번호판의 크기에 관계없이, 또한 잡음에 크게 영향을 받지 않으면서 번호판 추출이 정확하여 실시간 처리의 가능성을 제시하였을 뿐만 아니라 번호판 영역이 불투명하거나 불규칙한 조명 상태에서도 검출이 가능하였다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.67-69
• /
• 2021

Weather radar images can be used in a variety of ways because of their high visibility in terms of visuals. In other words it has the advantage of being able to grasp the flow of weather phenomena using not only the raw data of the weather radar, but also the change characteristics between consecutive images. In particular image processing techniques are gradually expanding in the field of meteorological research, and in the case of image data having high resolution such as weather radar images it is expected to produce useful information through a new approach called image processing techniques. In this study the weather phenomena flow was calculated as a vector from the change of the weather radar image according to time interval with the optical flow method, one of the image processing techniques. The characteristics of the weather phenomena to be analyzed were derived through vector analysis resolution suitable for the scale of weather, vector interpolation in regions where no radar echo exists, and the removal of relative flow vectors to distinguish the flow of specific weather and the entire atmosphere. Through this study, it is expected that not only the use of raw data of weather radar, but also the widening of the application area of weather radar, such as the use of unique characteristics of image data, and the active use of image processing techniques in the field of meteorology in the future.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.6 no.2
• /
• pp.46-64
• /
• 2003

The vector shaped object is generally very long (hundreds or thousands of kilometers) and very narrow (10-100 meters). Image mapping techniques and tools for these objects should be totally different from the traditional area-based targets. Acknowledging these unique characteristics of the vector shaped object, a motion picture mapping system has been developed by combining internet GIS technology with airborne video. In particular, integration between airborne video and digital maps took advantage of each component, and enabled the landscape structure to be visualized, interacted with and deployed all on the Web. The motion picture maps provided a completely new means for disseminating information for area-wide landscape in a visual and interactive manner to the general public while digital map with location information revealed successfully the major parameters that influence an area-wide spatial structure in the study area. The remote video approach breaks down the usual concept of image mapping in a conventional cartography. As a result, the research findings have established the new concept of 'internet airborne video mapping for vector shaped object', proposed as an initial aim of this paper. It would playa crucial role in improving the quality of public information service if the mapping system is operationally introduced into the Government since the highly user-friendly moving picture provides a completely new means for disseminating spatia) information for vector shaped object.

• Journal of KIISE:Information Networking
• /
• v.31 no.3
• /
• pp.298-304
• /
• 2004

In transmitting a compressed video bit-stream over Internet, a packet loss causes an error propagation in both spatial and temporal domain, which in turn leads to a severe degradation in image quality. In this paper, a new error concealment algorithm is proposed to repair damaged portions of the video frames in the receiver. Conventional BMA(Boundary Matching Algorithm) assumes that the pixels on the boundary of the missing block and its neighboring blocks are very similar, but has no consideration of edges t)r discontinuity across the boundary. In our approach, the edges are detected across the boundary of the lost or erroneous block. Once the edges are detected and the orientation of each edge is found, only the pixel difference along the expected edges across the boundary is measured instead of calculating differences between all adjacent pixels on the boundary. Therefore, the proposed approach needs very few computations and the experiment shows an improvement of the performance over the conventional BMA in terms of both subjective and objective quality of video sequences.