Optimal Camera Parameter Selection for State Estimation with Applications in Object Recognition

Abstract

In this paper we introduce a formalism for optimal camera parameter selection for iterative state estimation. We consider a framework based on Shannon’s information theory and select the camera parameters that maximize the mutual information, i.e. the information that the captured image conveys about the true state of the system. The technique explicitly takes into account the a priori probability governing the computation of the mutual information. Thus, a sequential decision process can be formed by treating the a posteriori probability at the current time step in the decision process as the a priori probability for the next time step. The convergence of the decision process can be proven.

We demonstrate the benefits of our approach using an active object recognition scenario. The results show that the sequential decision process outperforms a random strategy, both in the sense of recognition rate and number of views necessary to return a decision.