We develop a comprehensive description of the active inference framework, as proposed by Friston (2010), under a machine-learning compliant perspective. Stemming from a biological inspiration and the auto-encoding principles, a sketch of a cognitive architecture is proposed that should provide ways to implement estimation-oriented control policies under a POMDP perspective. Computer simulations...

Sensing three-dimensional shape is a central problem in the development of robot systems for autonomous navigation and manipulation. Stereo vision is an attractive approach to this problem in several applications; however, stereo algorithms still lack reliability and generality. We address these problems by modelling the stereo depth map as a discrete random field, by formulating the matching p...

Active learning (AL) is a paradigm where an active learner has to train a model (e.g., a classifier) which is in principle trained in a supervised way. In contrast to supervised learning, AL has to be done by means of a data set where a low fraction of samples is labeled or even with an initially unlabeled set of samples. To obtain labels for the unlabeled samples, the active learner has to ask...

Zoom lenses appear to fit very naturally into the framework of active vision — controlling a zoom lens allows an adjustment of the image, enabling either an analysis of a wide scene or a close look at a region or object of particular interest. However, their integration into vision systems is not without difficulty since zoom interacts insidiously with both low and high level processes. This th...

Abstract. Active cameras provide a mobile robot with the capability to fixate and track features over a wide field of view. However, their use emphasises serial attention focussing on a succession of scene features, raising the question of how this should be best achieved to provide localisation information. This paper describes a fully automatic system, able to detect, store and track suitable...

There are two main approaches to active vision for classification. The probabilistic approach employs a belief state concerning the class of the instance and selects actions to reduce uncertainty in the belief state. The situated approach directly learns a mapping from observations to actions (including the classification), driven by the performance on the classification task. One of the advant...

Recent demands in sophisticated mobile robots require many semi-autonomous or even autonomous operations, such as decision making, simultaneous localization and mapping, motion tracking and risk assessment, while operating in dynamic environments. Most of these capabilities depend highly on the quality of the input from the cameras mounted on the mobile platforms and require fast processing tim...

Machine vision is a powerful tool for process automation and control. However, the memory requirements and interface overheads of video buffers found in typical video grabbers are not easy to interface to low cost embedded systems. This paper describes an active vision interface that enables a low end microcontroller to use a standard video camera as an input for process control. The active vis...

We develop an architecture for reactive visual analysis of dynamic scenes. We specify a minimal set of system features based upon biological observations. We implement feature on a processing network based around an active stereo vision mechanism. Active rectification and mosaicing allows static stereo algorithms to operate on the active platform. Foveal zero disparity operations permit attende...

Multi-focal vision provides two or more vision devices with different fields of view and measurement accuracies. A main advantage of this concept is a flexible allocation of these sensor resources accounting for the current situational and task performance requirements. Particularly, vision devices with large fields of view and low accuracies can be used together. Thereby, a coarse overview of ...