TY - JOUR
T1 - Adaptive selection model for detecting zero-quantized discrete cosine transform coefficients in video coding
AU - Chen, Bo Jhih
AU - Tai, Shen-Chuan
AU - Wu, Yung Gi
PY - 2012/4/1
Y1 - 2012/4/1
N2 - An adaptive selection approach for advanced video coding that reduces the number of unnecessary discrete cosine transform (DCT), quantization (Q), inverse quantization (IQ), and inverse DCT (IDCT) computations is proposed. An adaptive selection approach is used to detect zero-quantized DCT (ZQDCT) coefficients since certain zero-quantized DCT coefficients of the prediction residual block are difficult to identify. The presented algorithm detects ZQDCT coefficients according to the corresponding frequency positions before implementing DCT. Therefore, redundant DCT, Q, IQ, and IDCT procedures are removed. To carry out the proposed algorithm, three sufficient conditions are assumed for deriving eight prediction modes, which are employed to select various types of DCT, Q, IQ, and IDCT implementations. The computed results indicate that the proposed algorithm is capable of reducing the number of DCT, Q, IQ, and IDCT computations compared to those required by related methods, while retaining the coding performance of the original encoder.
AB - An adaptive selection approach for advanced video coding that reduces the number of unnecessary discrete cosine transform (DCT), quantization (Q), inverse quantization (IQ), and inverse DCT (IDCT) computations is proposed. An adaptive selection approach is used to detect zero-quantized DCT (ZQDCT) coefficients since certain zero-quantized DCT coefficients of the prediction residual block are difficult to identify. The presented algorithm detects ZQDCT coefficients according to the corresponding frequency positions before implementing DCT. Therefore, redundant DCT, Q, IQ, and IDCT procedures are removed. To carry out the proposed algorithm, three sufficient conditions are assumed for deriving eight prediction modes, which are employed to select various types of DCT, Q, IQ, and IDCT implementations. The computed results indicate that the proposed algorithm is capable of reducing the number of DCT, Q, IQ, and IDCT computations compared to those required by related methods, while retaining the coding performance of the original encoder.
UR - http://www.scopus.com/inward/record.url?scp=84891283870&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84891283870&partnerID=8YFLogxK
U2 - 10.1117/1.OE.51.4.047404
DO - 10.1117/1.OE.51.4.047404
M3 - Article
AN - SCOPUS:84891283870
SN - 0091-3286
VL - 51
JO - Optical Engineering
JF - Optical Engineering
IS - 4
M1 - 047404
ER -