• 전자공학회논문지
• /
• 제51권1호
• /
• pp.107-115
• /
• 2014

영상의 저 비트 변환 기반의 비트 플레인 정합방법(Bit-Plane Matching : BPM)은 기존의 블록 정합 방법들과 비교해 계산량을 줄이고 간단한 하드웨어 구조 설계를 통해 블록 정합 결과를 획득할 수 있지만, 블록 정합의 정확도가 비교적 낮은 문제점을 가지고 있다. 본 논문에서는 기존의 BPM방법들과 비교해 블록 정합의 정확도를 증가시키면서 동시에 논리 연산으로 정합 결과를 계산할 수 있는 저 비트 변환 기반의 절대 오차합(Reduced-bit transform based Sum of Absolute Difference : R-SAD)을 이용한 블록 정합 알고리즘을 제안한다. 이 방법은 현재 영상과 참조영상을 각각 2-bit의 영상으로 변환하고, 2-bit의 4레벨에 대한 입출력 관계를 이용하여 진리표를 획득한다. 진리표는 Karnaugh map을 통해 간소화 되어 논리 연산으로 절대 오차를 계산할 수 있다. 제안된 방법의 성능 평가를 위한 움직임 보상(Motion Compensation) 실험에서, R-SAD는 기존의 블록 정합 방법들과 비교해 높은 정확도의 정합결과를 획득할 수 있었다.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 1996년도 학술대회
• /
• pp.217-220
• /
• 1996

This paper describes a new hierarchial block matching algorithm especially appropriate for a large search area. The proposed algorithm consists of higher level search for an initial motion vector estimate by using a new matching criterion over the evenly subsampled search points, and lower level search for the final motion vector refinement. In the higher level matching criterion, mean absolute differences at the search points (or motion vector candidates) similar to motion vectors of causally neighboring blocks, are weighted properly so that these points can have a higher chance to being selected. The proposed algorithm outperforms existing hierarchical block matching algorithms, and its computational regularity makes hardware implementation simple.

• 한국통신학회논문지
• /
• 제28권2C호
• /
• pp.121-128
• /
• 2003

본 논문에서는 스테레오 영상 압축 시 흔히 사용되는 블록 정합 기법의 문제점인 블록화 에러를 줄일 수 있는 객체기반 블록 정합 기법(object based block matching algorithm : OBMA)을 제안한다. 제안된 OBMA 기법에서는 영상의 경계 정보(edge map)를 추출한 후 블록내의 경계선 정보의 유무를 가지고 영상을 각각의 객체로 나누어 객체별로 변이를 예측함으로써 기존 블록정합 기법에서 발생하던 경계부분에서의 블록화 에러를 줄일 수 있었다. 모의 실험 결과 기존의 블록 정합(block matching) 기법보다 주관적 영상 품질은 현저히 개선하면서도 유사한 성능의 PSNR을 유지할 수 있었다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2000년도 ITC-CSCC -2
• /
• pp.843-846
• /
• 2000

An optical flow-based motion estimation algorithm is proposed for video coding. The algorithm uses block-matching motion estimation with an adaptive search region. The search region is computed from motion fields that are estimated based on the optical flow. The algorithm is based on the fact that true block-motion vectors have similar characteristics to optical flow vectors. Thereafter, the search region is computed using these optical flow vectors that include spatial relationships. In conventional block matching, the search region is fixed. In contrast, in the new method, the appropriate size and location of the search region are both decided by the proposed algorithm. The results obtained using test images show that the proposed algorithm can produce a significant improvement compared with previous block-matching algorithms.

• 한국멀티미디어학회논문지
• /
• 제14권10호
• /
• pp.1275-1286
• /
• 2011

This paper proposes a high speed systolic array architecture for full search block matching algorithm (FBMA). The pixels of the search area for a reference block are input only one time to find the matched candidate block and reused to compute the sum of absolute difference (SAD) for the adjacent candidate blocks. Each row of designed 2-dimensional systolic array compares the reference block with the adjacent blocks of the same row in search area. The lower rows of the designed array get the pixels from the upper row and compute the SAD with reusing the overlapped pixels of the candidate blocks within same column of the search area. This designed array has no data broadcasting and global paths. The comparison with existing architectures shows that this array is superior in terms of throughput through it requires a little more hardware.

• 전자공학회논문지S
• /
• 제34S권2호
• /
• pp.43-51
• /
• 1997

In this paper, we quantize the image data to simplify the systolic array architecture for block matching algorithm. As the number of bits for pixel data to be processed is reduced by quantization, one can simplify the hardware of systolic array. Especially, if the bit serial input is used, one can even more simplify the structure of processing element. First, we analize the effect of quantization to a block matching. then we show the structure of quantizer and processing element when bit serial input is used. The simulation results applied to standard images have shown that the proposed block matching method has less prediction error than the conventional high speed algorithm.

• 한국콘텐츠학회논문지
• /
• 제3권4호
• /
• pp.11-16
• /
• 2003

본 논문에서는 비디오코딩의 움직임 추정을 위한 빠른 블록 합 피라미드 알고리즘(FBSPA: Fast Block Sum Pyamid Algorithm)을 제안하였다. 제안된 알고리즘은 블록 합 피라미드 알고리즘, 블록 정합 움직임 추정을 위한 효율적인 다단계 연속 제거 알고리즘, 움직임 벡터 추정을 위한 빠른 알고리즘에 기초하고 있다. 제안된 알고리즘은 블록 합 피라미드알고리즘의 연산량을 감소시키기 위하여 블록 합 피라미드알고리즘에 부분 왜곡 제거 기법을 적용하였다. 제안된 방안이 블록 합 피라미드 알고리즘의 연산량을 최대 2.9% 감소시킬 수 있음을 실험을 통하여 확인하였다.

• 전자공학회논문지S
• /
• 제34S권9호
• /
• pp.77-83
• /
• 1997

We present a fast adaptive block matching algorithm using variable search area and subsampling to estimate motion vector more exactly. In the presented method, the block is classified into one of three motion categories: zero motion vector block, medium-motion bolck or high-motion block according to mean absolute difference of the block. By the simulation, the computation amount of the presented methoe comparable to three step search algorithm and new three step search algorithm. In the fast image sequence, the PSNR of our algorithm increased more than TSS and NTSS, because our algorithm estimated motion vector more accurately.

• Journal of Communications and Networks
• /
• 제12권1호
• /
• pp.5-10
• /
• 2010

This paper introduces a new fast motion estimation based on estimating a block matching error (i.e., sum of absolute difference (SAD)) between blocks which can eliminate an impossible candidate block much earlier than a conventional partial distortion elimination (PDE) scheme. The basic idea of the proposed scheme is based on predicting the total SAD of a candidate block using its partial SAD. In particular, in order to improve prediction accuracy and computational efficiency, a sub-sample based block matching and a selective pixel-based approaches are employed. In order to evaluate the proposed scheme, several baseline approaches are described and compared. The experimental results show that the proposed algorithm can reduce the computations by about 44% for motion estimation at the cost of 0.0005 dB quality degradation versus the general PDE algorithm.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 하계종합학술대회 논문집 Ⅳ
• /
• pp.2108-2111
• /
• 2003

This paper introduces a new local feature extraction method and image matching method for the localization and classification of targets. Proposed method is based on the block-by-block projection associated with directional pattern of blocks. Each pattern has its own eigen-vertors called as CEBs(Classified Eigen-Blocks). Also proposed block-based image matching method is robust to translation and occlusion. Performance of proposed feature extraction and matching method is verified by the face localization and FLIR-vehicle-image classification test.