Lucas Paletta

Appearance-based active object recognition

We present an efficient method within an active vision framework for recognizing objects which are ambiguous from certain viewpoints. The system is allowed to reposition the camera to capture additional views and, therefore, to improve the classification result obtained from a single view. The approach uses an appearance-based object representation, namely the parametric eigenspace, and augments it by probability distributions. This enables us to cope with possible variations in the input images due to errors in the pre-processing chain or changing imaging conditions. Furthermore, the use of probability distributions gives us a gauge to perform view planning. Multiple observations lead to a significant increase in recognition rate. Action planning is shown to be of great use in reducing the number of images necessary to achieve a certain recognition performance when compared to a random strategy. q 2000 Elsevier Science B.V. All rights reserved.

Active object recognition by view integration and reinforcement learning

Abstract A mobile agent with the task to classify its sensor pattern has to cope with ambiguous information. Active recognition of three-dimensional objects involves the observer in a search for discriminative evidence, e.g., by change of its viewpoint. This paper defines the recognition process as a sequential decision problem with the objective to disambiguate initial object hypotheses. Reinforcement learning provides then an efficient method to autonomously develop near-optimal decision strategies in terms of sensorimotor mappings. The proposed system learns object models from visual appearance and uses a radial basis function (RBF) network for a probabilistic interpretation of the two-dimensional views. The information gain in fusing successive object hypotheses provides a utility measure to reinforce actions leading to discriminative viewpoints. The system is verified in experiments with 16 objects and two degrees of freedom in sensor motion. Crucial improvements in performance are gained using the learned in contrast to random camera placements.

A comparison of probabilistic, possibilistic and evidence theoretic fusion schemes for active object recognition

Abstract.One major goal of active object recognition systems is to extract useful information from multiple measurements. We compare three frameworks for information fusion and view-planning using different uncertainty calculi: probability theory, possibility theory and Dempster-Shafer theory of evidence. The system dynamically repositions the camera to capture additional views in order to improve the classification result obtained from a single view. The active recognition problem can be tackled successfully by all the considered approaches with sometimes only slight differences in performance. Extensive experiments confirm that recognition rates can be improved considerably by performing active steps. Random selection of the next action is much less efficient than planning, both in recognition rate and in the average number of steps required for recognition. As long as the rate of wrong object-pose classifications stays low the probabilistic implementation always outperforms the other approaches. If the outlier rate increases averaging fusion schemes outperform conjunctive approaches for information integration. We use an appearance based object representation, namely the parametric eigenspace, but the planning algorithm is actually independent of the details of the specific object recognition environment.

Active Object Recognition in Parametric Eigenspace

We present an efficient method within an active vision framework for recognizing objects which are ambiguous from certain viewpoints. The system is allowed to reposition the camera to capture additional views and, therefore, to resolve the classification result obtained from a single view. The approach uses an appearance based object representation, namely the parametric eigenspace, and augments it by probability distributions. This captures possible variations in the input images due to errors in the pre-processing chain or the imaging system. Furthermore, the use of probability distributions gives us a gauge to view planning. View planning is shown to be of great use in reducing the number of images to be captured when compared to a random strategy.

Attention and Performance in Computational Vision

Attention in Object and Scene Recognition.- Distributed Control of Attention.- Inherent Limitations of Visual Search and the Role of Inner-Scene Similarity.- Attentive Object Detection Using an Information Theoretic Saliency Measure.- Architectures for Sequential Attention.- A Model of Object-Based Attention That Guides Active Visual Search to Behaviourally Relevant Locations.- Learning of Position-Invariant Object Representation Across Attention Shifts.- Combining Conspicuity Maps for hROIs Prediction.- Human Gaze Control in Real World Search.- Biologically Plausible Models for Attention.- The Computational Neuroscience of Visual Cognition: Attention, Memory and Reward.- Modeling Attention: From Computational Neuroscience to Computer Vision.- Towards a Biologically Plausible Active Visual Search Model.- Modeling Grouping Through Interactions Between Top-Down and Bottom-Up Processes: The Grouping and Selective Attention for Identification Model (G-SAIM).- TarzaNN : A General Purpose Neural Network Simulator for Visual Attention Modeling.- Applications of Attentive Vision.- Visual Attention for Object Recognition in Spatial 3D Data.- A Visual Attention-Based Approach for Automatic Landmark Selection and Recognition.- Biologically Motivated Visual Selective Attention for Face Localization.- Accumulative Computation Method for Motion Features Extraction in Active Selective Visual Attention.- Fast Detection of Frequent Change in Focus of Human Attention.

Visual object detection for autonomous sewer robots

The goal of the proposed detection system is to identify objects, e.g. inlets, in sewage pipes. A camera attached to an autonomous sewer robot provides images that are interpreted by an attention driven recognition module. Local appearances in the input image are represented in an environment specific description subspace extracted by principal component analysis. The object class posterior interpretation in terms of a radial basis function network constitutes an attention filter constraining further processing on receptive fields filter resolutions. Multiresolution decision fusion is the framework used to combine detection confidences to enhance robustness in the global classification. The vision system is evaluated in various experiments where it proves successful with respect to the local classification rate, to the generalization behavior in recognizing similar objects, and to detection that requires a minimum of positive falses.

Attention in Cognitive Systems. Theories and Systems from an Interdisciplinary Viewpoint

Embodiment of Attention.- The Embodied Dynamics of Emotion, Appraisal and Attention.- The Role of Attention in Creating a Cognitive System.- The Influence of the Body and Action on Spatial Attention.- Abstraction Level Regulation of Cognitive Processing Through Emotion-Based Attention Mechanisms.- Embodied Active Vision in Language Learning and Grounding.- Language Label Learning for Visual Concepts Discovered from Video Sequences.- Cognitive Control of Attention.- Learning to Attend - From Bottom-Up to Top-Down.- An Attentional System Combining Top-Down and Bottom-Up Influences.- The Selective Attention for Identification Model (SAIM): Simulating Visual Search in Natural Colour Images.- A Bayesian Approach to Attention Control and Concept Abstraction.- Modeling of Saliency and Visual Search.- An Information Theoretic Model of Saliency and Visual Search.- An Experimental Comparison of Three Guiding Principles for the Detection of Salient Image Locations: Stability, Complexity, and Discrimination.- A Proto-object Based Visual Attention Model.- Context Driven Focus of Attention for Object Detection.- Color Saliency and Inhibition Using Static and Dynamic Scenes in Region Based Visual Attention.- I See What You See: Eye Movements in Real-World Scenes Are Affected by Perceived Direction of Gaze.- Sequential Attention.- Selective Attention in the Learning of Viewpoint and Position Invariance.- Generating Sequence of Eye Fixations Using Decision-Theoretic Attention Model.- Reinforcement Learning for Decision Making in Sequential Visual Attention.- Biologically Inspired Framework for Learning and Abstract Representation of Attention Control.- Biological Aspects of Attention.- Modeling the Dynamics of Feature Binding During Object-Selective Attention.- The Spiking Search over Time and Space Model (sSoTS): Simulating Dual Task Experiments and the Temporal Dynamics of Preview Search.- On the Role of Dopamine in Cognitive Vision.- Differences and Interactions Between Cerebral Hemispheres When Processing Ambiguous Words.- Attention in Early Vision: Some Psychophysical Insights.- Auditory Gist Perception: An Alternative to Attentional Selection of Auditory Streams?.- Applications of Attentive Vision.- Simultaneous Robot Localization and Mapping Based on a Visual Attention System.- Autonomous Attentive Exploration in Search and Rescue Scenarios.- Attention-Based Landmark Selection in Autonomous Robotics.- Simulation and Formal Analysis of Visual Attention in Cognitive Systems.- Region-Oriented Visual Attention Framework for Activity Detection.- Autonomous Attentive Exploration in Search and Rescue Scenarios.

Attention Architectures for Machine Vision and Mobile Robots

ABSTRACT Computer vision systems that are applied for image understanding in real-world environments require the capability to focus operations on task relevant events in an ongoing input stream of visual information. Attentive systems must indirectly provide solutions to characteristic challenges in real-world processing, such as the complexity in input imagery and uncertainty in the acquired information. We address successful methodologies on saliency and feature selection, describe attentive systems with respect to object and scene recognition, and review saccadic interpretation under decision processes. In robotic systems, we understand attention embedded in the context of optimizing sensorimotor behavior and multisensor-based active perception. We present an overview on system architectures that play a crucial role in attentive robots, with emphasis on multimodal information fusion and humanoid robots.