M. Chahine, "Estimation of accelerated motion and occlusions from time-varying images, " Master's thesis, McGill University, Dept. Electr. Eng., July 1994. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Estimation and Compensation of Accelerated Motion for.. - Chahine, Konrad (1995) (1 citation) Self-citation (Chahine) (Correct)
....until convergence. Then, the last estimate b p is used as the new p, and the process is repeated. However, we have obtained better performance for p i = p i with on the fly update of p i (Gauss Seidel relaxation) This corresponds to one relaxation step between two updates of p i [6]. To improve the computational efficiency of the algorithm as well as the likelihood of convergence to the global minimum, a multiresolution approach is used. First, a pyramid of images is generated using Gaussian filters [12] Then, estimation is performed for the lowest resolution images and the ....
....image after low pass filtering and subsampling. To simulate motion of the rectangle at 1 4 pixel accuracy its intensities at consecutive time instants are computed via 4:1 subsampling with a variable spatial offset. This approach allows more realistic testing than in the case of pixel accuracy [6]. Figure 3 The accuracy of motion estimates is evaluated using the mean squared error (MSE) between components v x ; v y ; a x ; a y of the true motion field p and the estimated field b p in region R. R can be either the full estimation area (R 0 ) or the area of the moving rectangle (R 1 ) ....
[Article contains additional citation context not shown here]
M. Chahine, "Estimation of accelerated motion and occlusions from time-varying images, " Master's thesis, McGill University, Dept. Electr. Eng., July 1994.
Estimation of Trajectories for Accelerated Motion from.. - Chahine, al. (1994) (1 citation) Self-citation (Chahine) (Correct)
.... variance U i g (fl t i ) X k2I t [eg(x k i ; t k) Gamma i(x i ; t) 2 : 5) expressing the variability of intensity along c( x i ; t) with respect to the sample mean i i(x i ; t) 1 N X k2I t eg(x k i ; t k) eg(x k i ; t k) is the bicubic interpolated intensity [1] at time t k and position x k i : x k i = x i v(x i ; t)k a(x i ; t)k 2 = x i Delta k p i with Delta k = k 0 k 2 0 0 k 0 k 2 ; p i = Theta v x (x i ; t) v y (x i ; t) a x (x i ; t) a y (x i ; t) T : 5. OBJECTIVE FUNCTION Combining (3) and (4) MAP estimation (1) ....
....an algorithm to estimate velocity and acceleration in dynamic images. It has been demonstrated to give precise estimates for synthetic motion. Recently, we have applied the algorithm to images containing natural motion and we have evaluated the estimates using motion compensated interpolation [1]. We have obtained on average a 2 4 dB PSNR gain in the reconstruction error. Detailed account of those results will be given in a forthcoming publication. 9. ....
M. Chahine, "Estimation of accelerated motion and occlusions from time-varying images," Master's thesis, McGill University, Dept. Electr. Eng., July 1994.
On Motion Modeling and Estimation for Very Low Bit.. - Konrad, Mansouri.. (1995) (2 citations) (Correct)
No context found.
M. Chahine, "Estimation of accelerated motion and occlusions from time-varying images," Master's thesis, McGill University, Dept. Electr. Eng., July 1994.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC