Publication View

Evidential Reasoning for Object Recognition (2007)

Abstract
Evidential reasoning is the central problem of selective visual perception. It consists of gathering the evidence and integrating the information from successive fixations. Both data-driven and knowledge-driven influences in selecting the areas of interest and gathering the evidence are studied as well as their cooperative function. We use the Bayesian networks for knowledge representation and a decision theoretic approach in selecting where to shift the attention and for integrating the information from successive fixations. Depending upon the task the system decides what information to gather from the scale-space decomposition of an input image.

Publication details
Download http://citeseerx.ist.psu.edu/viewdoc/summary?doi=?doi=10.1.1.15.6981
Source http://www.caip.rutgers.edu/~marsic/Publications/annie94.pdf
Contributors CiteSeerX
Repository CiteSeerX - Scientific Literature Digital Library and Search Engine (United States)
Keywords object recognition, selective perception, Bayesian networks, multiresolution analysis
Type text
Language English
Relation 10.1.1.30.1227