TY - JOUR
T1 - Implementation of Motion Estimation Algorithms in Multi-scalability to Provide High-efficiency Video Coding
AU - Purwadi, Agus
AU - Suwadi,
AU - Wirawan,
AU - Prakasa, Esa
N1 - Publisher Copyright:
© (2024), (Intelligent Network and Systems Society). All rights reserved.
PY - 2024
Y1 - 2024
N2 - Motion estimation involves determining the direction of an object’s movement, which is crucial in video coding during the transmission process. The motion vector can indicate the shift point between the currently processed frame and the frame used as a processing reference. The sum-of-absolute-difference (SAD) block-matching algorithm relies heavily on estimating object movement. In this research, we propose the integration of three-step search (TSS) and full search (FS) methods in multi-scalable video transmission using high-efficiency video coding (HEVC), applying three scalabilities—spatial, signal-to-noise ratio (SNR), and temporal. This integration aims to increase efficiency and improve quality by employing a block-matching algorithm. With this design, we evaluate the performance of the TSS and FS methods in multi-scalable video coding, obtaining the video frame quality with peak SNR (PSNR) and bit rate efficiency. From the results of experiments using video tests in Standard Definition (SD), Common Intermediate Format (CIF), and High- Definition (HD) formats, the FS algorithm has a total Bjontegaard delta (BD)-PSNR value of 0 dB and an efficiency of 62.4%, while the TSS achieves a total BD-PSNR value of 0.8 dB and an efficiency of 23.6%. Meanwhile, the optimal PSNR and bit rate for the multi-scalability average were found with the FS algorithm enhancement layer and the TSS algorithm enhancement layer.
AB - Motion estimation involves determining the direction of an object’s movement, which is crucial in video coding during the transmission process. The motion vector can indicate the shift point between the currently processed frame and the frame used as a processing reference. The sum-of-absolute-difference (SAD) block-matching algorithm relies heavily on estimating object movement. In this research, we propose the integration of three-step search (TSS) and full search (FS) methods in multi-scalable video transmission using high-efficiency video coding (HEVC), applying three scalabilities—spatial, signal-to-noise ratio (SNR), and temporal. This integration aims to increase efficiency and improve quality by employing a block-matching algorithm. With this design, we evaluate the performance of the TSS and FS methods in multi-scalable video coding, obtaining the video frame quality with peak SNR (PSNR) and bit rate efficiency. From the results of experiments using video tests in Standard Definition (SD), Common Intermediate Format (CIF), and High- Definition (HD) formats, the FS algorithm has a total Bjontegaard delta (BD)-PSNR value of 0 dB and an efficiency of 62.4%, while the TSS achieves a total BD-PSNR value of 0.8 dB and an efficiency of 23.6%. Meanwhile, the optimal PSNR and bit rate for the multi-scalability average were found with the FS algorithm enhancement layer and the TSS algorithm enhancement layer.
KW - FS
KW - HEVC
KW - SAD
KW - SVC
KW - TSS
UR - http://www.scopus.com/inward/record.url?scp=85199791263&partnerID=8YFLogxK
U2 - 10.22266/IJIES2024.0831.76
DO - 10.22266/IJIES2024.0831.76
M3 - Article
AN - SCOPUS:85199791263
SN - 2185-310X
VL - 17
SP - 1005
EP - 1028
JO - International Journal of Intelligent Engineering and Systems
JF - International Journal of Intelligent Engineering and Systems
IS - 4
ER -