• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.1C
• /
• pp.34-44
• /
• 2008

In this paper, a new rate-distortion estimation method is proposed for fast rate-distortion optimized(RDO) mode decision in H.264 encoder. A bit-rate and distortion should be calculated by encoding and reconstruction of each macroblock mode in order to select the optimal RDO mode. To reduce the computations for getting them, a bit-rate and distortion is efficiently estimated in DCT domain. Simulation results show that the proposed method achieves computational saving about 81% of RDO mode decision and saves about 26% of total encoding time without noticeable degradation in coding performance.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2013.11a
• /
• pp.172-173
• /
• 2013

최근 표준화가 완료된 HEVO(High Efficiency Video Coding)에서는 계층적 구조를 갖는 가변블록의 크기를 사용하고 재귀적으로 부호화를 수행사여, 최적의 부호화단위(CU: Coding Unit) 분할 구조와 예측단위(PU: Prediction Unit)를 결정함으로써 높은 부호화 효율을 얻을 수 있는 반면 부호화 복잡도가 증가하는 문제가 있다. 본 논문에서는 부호화기의 복잡도를 감소시키기 위한 고속 부호화 알고리즘으로 고속 모드 결정 기법을 제안한다. 제안기법은 상위 깊이(CU: Coding Unit)의 최적 모드와 부호화 율-왜곡 비용을 이용해서 현재 깊이 CU에서의 특정 모드의 율-왜곡 비용 계산을 생략함으로써 PU 탐색을 조기 종료한다. 즉, 상위 깊이 CU의 조건에 따라 화면간 예측 모드의 일부 또는 화면내 예측을 수행하지 않는다. 실험결과 제안기법은 HM 12.0대비 0.2%의 비트 증가에 22.9%의 계산시간 감소 효과를 얻을 수 있음을 확인하였다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.8 no.6
• /
• pp.1121-1125
• /
• 2004

This paper is studied the fast block mode decision algorithm for H.264/AVC. The fast block mode decision algorithm is consist of block range decision and merge algorithm. The block range decision algorithm classifies the block over 8$\times$8 size or below for 16$\times$16 macroblock to decide the size and type of sub blocks. The block over 8$\times$8 size is divided into the blocks of 16$\times$8, 8$\times$16 and 16$\times$16 size using merging algorithm which is considered MVD(motion vector difference) of 8$\times$8 block. The sub block range decision reduces encoding arithmetic amount by 48.25% on the average more than the case not using block range decision.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.43 no.3 s.309
• /
• pp.43-53
• /
• 2006

As wireless network technology has advanced, demands for multimedia contents through mobile environment have tendered to upward. Since network situation is changing every moment and types of user terminals are diverse, it is difficult for a content provider to consider network situation and type of user terminal to provide multimedia contents. As one solution, transcoding techniques have been proposed, but those have much complexity. In this paper, in order to reduce computational complexity, we propose a fast mode decision using input modes, motion vectors, and residual energies which are obtained from input bitstream for 2:1 down-scaling spatial transcoding application. The proposed method reduces processing time in mode decision by restricting possible mode types based on input information. Experimental results show that the proposed method achieves about 2.66 times improvement in encoding time compared to the normal encoding process while the PSNR is degraded by about 0.04dB, and bit-rate is increased by 1.6%.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2010.11a
• /
• pp.96-99
• /
• 2010

H.264/AVC 부호화 표준은 영상의 특성을 반영하기 위해 $16{\times}16$부터 $4{\times}4$ 크기의 가변적인 블록을 이용하여 부호화 효율을 높인다. 하지만 이로 인해 부호기의 복잡도가 증가된다. 부호기 복잡도를 증가시키는 여러 요인 중, H.264/AVC의 모드 결정은 부호기의 복잡도를 증가시키는 주요인이다. 본 논문에서는 IPPP구조에서 비트율 왜곡값을 이용하여 고속으로 매크로블록의 모드를 결정하는 방법을 제안한다. 인트라 화면에서의 인트라 $4{\times}4$, 인트라 $16{\times}16$의 비트율 왜곡 평균값으로 영상에 적응적인 최대 임계값과 최소 임계값을 결정한다. 다음, $16{\times}16$, $16{\times}8$, $8{\times}16$ 인터 모드의 비트율 왜곡값이 최대 임계값과 최소 임계값으로 분할한 범위 중 어느 곳에 해당하는지를 살펴보고, 이에 따라 인트라 모드 결정 단계를 선택적으로 결정한다. 제안하는 알고리즘은 기존의 H.264/AVC에 비해 부호화 효율의 큰 감소 없이 평균 27.42%의 부호화 시간을 감소시켰다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.7
• /
• pp.1589-1596
• /
• 2010

In this paper, we propose the method that reduce computational complexity for intermode decision using CBP(coded block pattern) and coded information of colocated-MB(macro block). Proposed method classifies MB into best-CBP and normal-CBP according to the characteristics of CBP. On best-CBP, it eliminates the computation for $8{\times}8$ mode on intermode decision process because the probability for SKIP mode and M-Type mode is 96.3% statistically. On normal-CBP, it selectively eliminates the amount of computation for bit-rate distortion cost, because it uses coded information of colocated-MB and motion vector cost in deciding SKIP mode and M-Type mode. The simulation results show that the proposed method reduces total coding time to 58.44% in average, and is effective in reducing computational burden in videos with little motion.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2015.11a
• /
• pp.143-144
• /
• 2015

최신 동영상 표준 코덱인 High Efficiency Video Coding (HEVC)는 기존의 H.264/AVC 보다 동일 화질 대비 최대 약 2배의 압축 성능을 보여준다. 이러한 성능을 얻기 위해 HEVC에는 다양한 압축 기술이 적용되었다. 그 예로, H.264/AVC에서는 인트라 예측 모드에서 9가지 예측 모드만을 사용한 반면 HEVC에서는 35가지의 모드를 이용해 화면 내 예측을 시행한다. HEVC에 적용된 다양한 기술들에 의해 부호화 복잡도가 증가하였고 복잡도를 줄이기 위해 다양한 고속 알고리즘이 연구되고 있다. 본 논문에서는 스크린 콘텐츠 영상 부호화에 적합한 고속 인트라 예측 알고리즘을 제안한다. 스크린 콘텐츠 영상이란 카메라를 이용해 촬영된 자연계 영상이 아닌 mobile phone, 방송 장비, 기타 전자 기기 등 컴퓨터 기술에 의해 생성되는 영상을 의미한다. 스크린 컨텐츠 영상은 자연계 영상과 달리 색의 변화가 전혀 없는 단순한 영역을 갖는 특성이 있다. 이러한 스크린 콘텐츠 영상의 특성을 반영하는 고속 알고리즘을 HEVC Test Model (HM) 16.6에 적용하였고, 스크린 컨텐츠 영상에서 25%의 속도 향상 결과를 얻을 수 있었다.

• Journal of Broadcast Engineering
• /
• v.14 no.1
• /
• pp.28-35
• /
• 2009

H.264/AVC, a recently developed video compression standard, is used for various applications because of its high coding efficiency. Variable block mode plays important role in the high coding efficiency of H.264/AVC but involves significant computations to select the optimal mode. In this paper, a fast mode decision method for H.264/AVC P slices is presented. To reduce computations for mode decision, the proposed mode decision method skips the mode decision processes for small partition modes using distortions of SKIP mode and intra16x16 mode. The experimental results show that the proposed method can reduce encoding time up to 66.41% while maintaining compression efficiency.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.56 no.10
• /
• pp.1868-1871
• /
• 2007

This paper proposes a novel scheme to significantly reduce mode decision time by predicting optimal mode candidates. Unlike previous fast mode decision algorithms computing RDcost in a pre-defined mode order, the proposed scheme predicts optimal mode candidates and calculates their RDcosts first, increasing the possibility to satisfy early-exit conditions sooner, resulting in fast mode decision. This H.264 mode decision time reduction enables small computing power mobile devices to handle H.264 encoding effectively. Extensive simulations show that, when compared with JM10.2, AMD and LCIMS, the proposed scheme boosts H.264 encoding speed by up to 575% with a reasonable image quality degradation.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.45 no.6
• /
• pp.95-107
• /
• 2008

In this paper, we propose a fast mode decision method for temporal and spatial scalability to reduce computational complexity of mode decision that used to be computationally one of the most intensive processes of the MPEG-4 AVC|H.264 SE(Scalable Extension) encoding. For temporal scalability, we propose an early skip method and MHM(mode history map) method. The early skip method confines macroblock modes of backward and forward frames within selected a few candidates. The MHM method utilizes stored information of frames inside a GOP of lower levels for the decision of MHM at higher level. For the spatial scalability, we propose the method that uses a candidate mode according to the MHM method and adds the BL_mode as candidates. The proposed scheme reduces the number of candidate modes to reduce computational complexity in mode decision. The proposed scheme reduces total encoding time by about 52% for temporal scalability and 47% for spatial scalability without significant loss of RD performance.