• 한국정보과학회논문지:소프트웨어및응용
• /
• 제33권1호
• /
• pp.95-104
• /
• 2006

Gradient vector flow(GVF) snake 또는 active contour 모델은 영상 분할에서 훌륭한 성능을 보여준다. 그러나 기존의 snake 모델에는 제한된 캡쳐 영역과 요면으로의 느린 진행과 같은 문제점들이 존재한다. 본 논문은 주변의 필드로부터 외부장(external force field)을 확장시키고 변형된 평탄화기법을 이용하여 확장된 필드를 정규화 함으로서 GVF snake 모델의 성능을 개선시키는 새로운 방법을 제시한다. 시뮬레이션을 위해 사용된 U자 모양 이미지에서의 결과는 제안된 방법이 좀 더 큰 캡쳐 영역을 갖고 기존의 GVF snake 모델에 비하여 요면으로 빠르게 진행하는 것이 가능함을 보여준다.

• 한국멀티미디어학회논문지
• /
• 제19권6호
• /
• pp.1003-1013
• /
• 2016

This paper proposes a framework of superpixel-based vehicle detection method using plane normal vector in disparity space. We utilize two common factors for detecting vehicles: Hypothesis Generation (HG) and Hypothesis Verification (HV). At the stage of HG, we set the regions of interest (ROI) by estimating the lane, and track them to reduce computational cost of the overall processes. The image is then divided into compact superpixels, each of which is viewed as a plane composed of the normal vector in disparity space. After that, the representative normal vector is computed at a superpixel-level, which alleviates the well-known problems of conventional color-based and depth-based approaches. Based on the assumption that the central-bottom of the input image is always on the navigable region, the road and obstacle candidates are simultaneously extracted by the plane normal vectors obtained from K-means algorithm. At the stage of HV, the separated obstacle candidates are verified by employing HOG and SVM as for a feature and classifying function, respectively. To achieve this, we trained SVM classifier by HOG features of KITTI training dataset. The experimental results demonstrate that the proposed vehicle detection system outperforms the conventional HOG-based methods qualitatively and quantitatively.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2015년도 추계학술대회
• /
• pp.101-104
• /
• 2015

영상에서 이전 프레임과 현재 프레임을 비교하여 얼마나 움직였는지를 나타내는 모션 벡터(motion vector)라 한다. 이러한 모션 벡터를 이용하여 현재 프레임의 이미지안의 객체들이 이전프레임의 이미지 객체가 있었던 위치로 보정을 한다면 손 떨림이나 카메라의 흔들림에서 오는 작은 떨림을 보정할 수 있다. 본 논문에서는 모션 벡터를 추출하는 방법인 SAD(Sum of Absolute Difference) 방정식을 이용한 블록 정합 매칭 알고리즘, 위상상관을 이용한 매칭, 특징점을 이용한 매칭, Bitplane을 이용한 블록매칭의 성능을 비교하였다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 하계종합학술대회 논문집 Ⅳ
• /
• pp.1747-1750
• /
• 2003

This paper proposes a multilevel vector error diffusion for fast and accurate color reproduction. Proposed method considered both hue angle and Euclidean distance during the multilevel vector error diffusion procedure to improve time complexity and output image quality In the error diffusion process, it can be determined whether error is diffused or not by comparing the vector norm and lightness value between original vector and error corrected vector of neighborhood pixels. For adaptive selection of output patch, this paper computes chroma value of error corrected vector and compares the hue angle between error corrected input vector and 64 primary color vectors.

• Journal of Electrical Engineering and information Science
• /
• 제3권3호
• /
• pp.330-335
• /
• 1998

In this paper, a new light-value approaching method to enhance a color image by excluding the effect of incident illumination is proposed. The method uses the fundamental vector transformation in which an estimated color of illumination is rotated to the white color of natural daylight. Then the transformed red, green, and blue values of each pixel are nonlinearly mapped into the 8-bit values to enhance intensity and saturation in the dark portion of the image. The proposed algorithm can produce the enhanced color image fast and efficiently without any space conversion or noticeable distortion.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2018년도 추계학술대회
• /
• pp.536-537
• /
• 2018

본 논문은 개와 고양이에 해당하는 애완동물 영상에서 세부 분류인 동물의 종을 인식하는 것을 목표로 한다. 영상의 세부 분류 인식에 대한 연구는 계속적으로 발전하고 있지만, 다형성의 성질을 갖는 동물에 대한 객체인식 연구는 더디게 진행되고 있다. 본 논문에서는 객체 분리를 위해 Grab-cut 알고리즘을 이용하고, 영상 인코딩을 위해 Fisher Vector를 이용하여 더 높은 동물 객체인식 방법을 제안한다.

• 한국컴퓨터정보학회논문지
• /
• 제23권10호
• /
• pp.95-101
• /
• 2018

In this paper, we propose an image-based synthesis method that can effectively represent the spark effect in fire simulation. We use the real flame image or animated image as inputs and perform the following steps : 1) extract feature vectors from the image, 2) calculate artificial buoyancy, and 3) generate and advect spark textures. We detect the edge from images and then calculate the feature vectors to calculate the buoyancy. In the next step, we compute the high-quality buoyancy vector field by integrating the two-dimensional feature vector and the fluid equation. Finally, the spark texture is advect by buoyancy field. As a result, our method is performed much faster than the previous approach and high-quality results can be obtained easily and stably.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2007년도 하계종합학술대회 논문집
• /
• pp.379-380
• /
• 2007

In this paper, we propose the fast hi-linear interpolation method for SR reconstruction. This method reconstructs the HR image rapidity by considering motion vector information for each LR input image. And its calculation used normalized deviation of image data. As using the motion vector information which is obtained at registration error checking process, this proposed can be achieved the fast and simple SR reconstructed image.

• 한국통신학회논문지
• /
• 제22권3호
• /
• pp.469-480
• /
• 1997

For moving image coding, the variable size of region coding based on local motion is more efficient than fixed size of region coding. It can be applied well to complex motions and is more stable for wide motions because images are segmented according to local motions. In this paper, new image coding method using the segmentation of motion vectors is proposed. First, motion vector field is smoothed by filtering and segmented by smoothed motion vectors. The region growing method is used for decomposition of regions, and merging of regions is decided by motion vector and prediction errors of the region. Edge of regions is excluded because of the correlation of image, and neighbor motion vectors are used evaluation of current block and construction of region. The results of computer simulation show the proposed method is superior than the existing methods in aspect of coding efficiency.

• 한국정보처리학회논문지
• /
• 제7권3호
• /
• pp.1013-1020
• /
• 2000

In this paper, we propose an algorithm which segments hierarchically blocks of image using discrete cosine transform(DCT) and execute finite-state vector quantization (FSVQ) for each block. Using DCT coefficient feature, image is segmented hierarchically to large smooth block and small edge block, then the block hierarchy informations are transmitted. The codebooks are respectively constructed for each hierarchical blocks, the encoder transmits codeword index using FSVQ for reducing encoded bit with hierarchical segmentation. Compared with side match VQ(SMVQ) and hierarchical FSVQ(HFSVQ) algorithm, about Zelda and Boat image, the new algorithm shows better picture quality with 1.97dB and 2.85 dB difference as to SMVQ, 1.78dB and 1.85dB diffences as to HFSVQ respectively.