Rao RPN, Ballard DH. Kalman filter model of the visual cortex. Neural Computation 1997; 9(4): 721-- 763  Home/Search   Document Not in Database   Summary   Related Articles   Check

....deals with a relaxation equation that will be called the DCR dynamics. Similar dynamical equations have been suggested to form sparse representation [22] These equations differ from the DCR equation since these latter include possible losses. Moreover, the predictive Kalman filter approach [23] has common points with the DCR equations since the Kalman filter approach also develops internal activities that minimizes the difference between the input and the reconstructed input. With regard to the reconstruction architecture but not the reconstruction dynamics a related architecture ....

R. P. N. Rao and D. H. Ballard. Kalman filter model of the visual cortex. Neural Computation, 1997. In press.

....structure that describes the flow field and thus the structure of temporal development will also be called temporal association. Since temporal associations are the means of prediction, this point of the model becomes similar to the predictive top down Kalman filtering structure of Rao and Ballard [30]. Beyond the dissimilarities of the respective architectures the main difference between the two models is that our starting point is a general control architecture that can be modified to a compression reconstruction model and then to a layered structure that resembles the basic circuit of ....

R. P. N. Rao and D. H. Ballard. Kalman filter model of the visual cortex. Neural Computation, 1997. In press.

No context found.

Rao RPN, Ballard DH. Kalman filter model of the visual cortex. Neural Computation 1997; 9(4): 721-- 763

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