Academia.eduAcademia.edu

Figure 1 – uploaded by ASSCOCIATE PROFESSOR TS. DR. HAMIDAH JANTAN JANTAN

See full PDFdownloadDownload figure
Dcomp and Dpsnr Comparison of the MDS and MARPS Algorithms for all Motion Types Figure 1: Performance Comparisons of DS and MDS Algorithms in term of Number of Computations of Foreman Sequence

Figure 1 Dcomp and Dpsnr Comparison of the MDS and MARPS Algorithms for all Motion Types Figure 1: Performance Comparisons of DS and MDS Algorithms in term of Number of Computations of Foreman Sequence

Related Figures (4)

Performance of Block Matching Algorithms in terms of no. of Computations and PSNR of High Motion Type Performance of Block Matching Algorithms in terms of no. of Computations and PSNR of Medium Motion Type
Performance of Block Matching Algorithms in terms of no. of Computations and PSNR of High Motion Type Performance of Block Matching Algorithms in terms of no. of Computations and PSNR of Medium Motion Type
Figure 2: Performance Comparisons of ARPS and MARPS Algorithms in term of Number of Computations of Claire Sequence
Figure 2: Performance Comparisons of ARPS and MARPS Algorithms in term of Number of Computations of Claire Sequence
Performance of Block Matching Algorithms in terms of no. of Computations and PSNR of Low Motion Type Table 4 shows the percentage degradation of the number of computations and PSNR values of the proposed MDS and MARPS algorithms as compared to those of DS and ARPS algorithms respectively for all motion type (H, M and L represent for high, medium and low motion type respectively). The sign (-) in Dcomp and Dpsnr indicate a loss values in their performances. MDS has reduced the number of computations by 7.56% for high motion sequence, 0.16% for medium motion sequence and 0.40% for low motion sequence. MARPS is considered as the fastest algorithm because it has greatly reduced the number of computations by 19.32%, 18.76% and 17.96% for low, medium and high motion sequence respectively. MDS and MARPS also provide good PSNR values with small degradation ratio for all motion types.
Performance of Block Matching Algorithms in terms of no. of Computations and PSNR of Low Motion Type Table 4 shows the percentage degradation of the number of computations and PSNR values of the proposed MDS and MARPS algorithms as compared to those of DS and ARPS algorithms respectively for all motion type (H, M and L represent for high, medium and low motion type respectively). The sign (-) in Dcomp and Dpsnr indicate a loss values in their performances. MDS has reduced the number of computations by 7.56% for high motion sequence, 0.16% for medium motion sequence and 0.40% for low motion sequence. MARPS is considered as the fastest algorithm because it has greatly reduced the number of computations by 19.32%, 18.76% and 17.96% for low, medium and high motion sequence respectively. MDS and MARPS also provide good PSNR values with small degradation ratio for all motion types.
Table 3
Table 3
Related topics:
EngineeringComputer ScienceMotion estimationBlock Matching Algorithm
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved