2 citations found. Retrieving documents...
S. X. Ju, Estimating Image Motion in Layers: The Skin and Bones Model, Ph.D. thesis, University of Toronto, 1999.

 Home/Search   Document Details and Download   Summary   Related Articles   Check  

This paper is cited in the following contexts:
Robustly Estimating Changes in Image Appearance - Black, Fleet, Yacoob (2000)   (17 citations)  (Correct)

....estimation [2, 28, 43 45] One issue with layered models concerns the estimation of the appropriate number of layers. A number of authors have used a minimum description length criterion to strike a balance between accurate encoding of the motion and the number of layers needed to represent it [2, 15, 27]. ROBUST ESTIMATION OF APPEARANCE CHANGE 13 With parameterized models, the estimation of the motion of a given surface may be adversely affected by distant, and quite unrelated, image points. These distant motions can act as leverage points [35] that pull the solution away from the desired ....

....may be adversely affected by distant, and quite unrelated, image points. These distant motions can act as leverage points [35] that pull the solution away from the desired local motion. A spatial smoothness constraint can be added to the computation of the weights that assign pixels to layers [27, 44]. This may reduce the effect of leverage points by encouraging layers to have spatially coherent support. Specialized spatial models. Parameterized approaches may perform poorly when the spatial variation of the image motion is more complex than a low order polynomial. To handle complex motions ....

S. X. Ju, Estimating Image Motion in Layers: The Skin and Bones Model, Ph.D. thesis, University of Toronto, 1999.


Robustly Estimating Changes in Image Appearance - Black, Fleet, al. (2000)   (17 citations)  (Correct)

....One issue with layered models concerns the estimation of the appropriate number of layers. A number of authors 4 BLACK, FLEET AND YACOOB have used a minimum description length criterion to strike a balance between accurate encoding of the motion and the number of layers needed to represent it [2, 15, 27]. With parameterized models, the estimation of the motion of a given surface may be adversely affected by distant, and quite unrelated, image points. These distant motions can act as leverage points [35] that pull the solution away from the desired local motion. A spatial smoothness constraint ....

....may be adversely affected by distant, and quite unrelated, image points. These distant motions can act as leverage points [35] that pull the solution away from the desired local motion. A spatial smoothness constraint can be added to the computation of the weights that assign pixels to layers [27, 44]. This may reduce the effect of leverage points by encouraging layers to have spatially coherent support. Specialized Spatial Models Parameterized approaches may perform poorly when the spatial variation of the image motion is more complex than a low order polynomial. To handle complex motions ....

S. X. Ju. Estimating image motion in layers: The Skin and Bones model. PhD thesis, University of Toronto, 1999.

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