Earticle

Integer Discrete Cosine Transform (DCT) is among the techniques used to improve the performance of the H.264/AVC Standard. All the profiles in the H.264/AVC standard support 4×4 integer DCT and the high profiles of this standard support 8×8 integer DCT as well as the 4×4 integer DCT. Various hardware realizations have been proposed for forward and inverse integer DCT in the literature because they are among the computational intensive units in the H.264/AVC standard. In this paper we propose a unified pipelined architecture to realize of the entire forward and inverse DCTs as well as the Hadamard transforms in the H.264/AVC encoder. The synthesis results indicate that our architecture achieves higher clock rate and relatively lower gate count compared to the other published architectures that realize only a number of the transforms in the H.264/AVC encoder.

DCT는 효율적인 에너지의 집중과 공간적 중복성을 제거하는 장점을 가지고 있어 영상압축 또는 비디오 신호처리등 다양한 분야에서 널리 사용되고 있다. 하지만 DCT는 많은 연산량의 문제를 가지고 있다. Loeffler에 의해 제안된 1차원 8-point DCT는 최소 11개의 곱셈과 29개의 덧셈을 이용하여 DCT 계수를 얻게 된다. 정수형 binDCT는 시스템의 복잡도를 증가시키는 곱셈 연산을 리프팅 구조를 이용해 덧셈과 지연으로 전환하여 DCT 계수를 구하게 되고 실수형 계수를 정수형의 계수로 바꾸어서 처리함으로써 시스템의 복잡도를 줄일 수 있다. DCT를 이용한 스케일러블 영상압축 알고리즘은 EZW[6], SPIHT와 같이 원하는 비트율까지 압축을 수행하게 되는데 다해상도 변환을 사용하는 경우에 비해 복잡도가 낮으면서도 높은 성능의 영상을 얻을 수 있는 장점이 있다.

The DCT has merit such effects of energy compression and removing redundancy that we found wide applications in image compression and video signal processing and other field. But DCT has high complexity. The theoretical lower bound on the number of multiplications required for the 1-D 8-pont DCT by loeffler has been proven to be 11 multiplications and 29 additions. The integer binDCT exchanges multiplication with addition and delay by using lifting scheme. Also, the integer binDCT reduce the complexity of the system by processing of exchanging floating point coefficients with integer coefficients. The scalable image compression using integer DCT accomplishes the compression on variable bit rate similar to EZW and SPIHT. It has the merit of lower complexity and higher performance than using multi-resolution transform.

The system for the real time face detection is described in this paper. For face verification, support vector machine (SVM) was utilized. Although SVM performs quit well, SVM has a drawback that the computational cost is high because all pixels in a mask are used as an input feature vector of SVM. To resolve this drawback, a method to reduce the dimension of feature vectors using the integer DCT was proposed. Also for the real time face detection applications, low-complexity methods for face candidate detection in a gray image were used. As a result, the accurate face detection was performed in real time.

This paper proposes a novel algorithm for hybrid lossless audio coding, which employs integer discrete cosine transform. The proposed algorithm divides the input signal into frames of a proper length, decorrelates the framed data using the integer DCT and finally entropy-codes the frame data. In particular, the adaptive Golomb-Rice coding method used for the entropy coding selects an optimal option which gives the best compression efficiency. Since the proposed algorithm uses integer operations, it significantly improves the computation speed in comparison with an algorithm using real or floating-point operations. When the coding algorithm is implemented in hardware, the system complexity as well as the power consumption is remarkably reduced. Finally, because each frame is independently coded and is byte-aligned with respect to the frame header, it is convenient to move, search, and edit the coded data.

본 논문에서는 기존의 H.264/AVC의 spatial 영역에서 Intra prediction 기법과 달리 H.264/AVC에서 사용하는 Integer DCT 영역에서 Intra prediction 기법을 제안한다. 이를 위하여 Integer DCT 영역에서 Intra prediction을 수행하는 모든 과정을 matrix multiplication으로 표현하여 Intra prediction을 수행하는 matrix를 유도한다. Intra prediction을 수행하는 matrix를 각 모드에 알맞게 설계하고, 이 matrix를 Integer DCT 영역에서 사용할 수 있도록 orthogonal한 Integer matrix를 설계한다. 실험을 통하여 제안한 Integer DCT 영역에서 Intra prediction 기법이 기존의 H.264/AVC의 spatial 영역에서 intra prediction 기법과 성능이 동일하면서 어떻게 matrix multiplication에 연산들을 포함시켜서 단순화 할 수 있는지를 보여주겠다. 또한 H.264/AVC에서 제공하는 intra prediction 각 모드에 대해 계산상 복잡도를 분석하였다.

This paper is studied the high speed processing moving picture encodec to compress and encode a moving picture by real time. This is used the new motion vector search algorithm with smallest search point in H.263 encodec, and is applied the integer DCT for the encodec by converting a moving picture. The integer DCT behaves DCT by the addition operation of the integer using WHT and a integer lifting than conventional DCT that needs the multiplication operation of a floating point number. Therefore, the integer DCT can reduce the operation amount than basis DCT with having an equal PSNR because the multiplication operation of a floating point number does not need.