R. Szeliski, "Regularization uses fractal priors;' Proc. AAAI-87, Seattle, pp. 749-754, 1987.  Home/Search   Document Not in Database   Summary   Related Articles

....later. Finally, the prior model can be used to embed prior knowledge about the scene. For the iconic method, lbr example, smoothness constraints requiring nearby image points to have similar disparity can be modeled eas ily by off diagonal elements of the inverse of the prior covariance matrix P0 [29]. Our algorithm incorporates Our actual implementation uses inverse depth (called disparity b See section 4. 212 Matthies, Kanade, Saeliski this knowledge as part of a smoothing operation that follows the state update stage. Similar concepts may be applicable to modeling figural continuity ....

....within our framework. An interesting issue we have not explored is the propagation of detected discontinuities between frames. The smoothing stage can be viewed as the part of the Kalman filtering algorithm that incorporates prior knowledge about the smoothness of the disparity map. As shown in [29], a regularization based smoother is equivalent to a prior model with a correlation function defined by the degree of the stabilizing spline (e.g. membrane or thin plate) In terms of table l, this means that the prior covariance matrix P0 is nondiagonal. The resulting posterior covariance matrix ....

R. Szeliski, "Regularization uses fractal priors;' Proc. AAAI-87, Seattle, pp. 749-754, 1987.

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