Q-learning of sequential attention for visual object recognition from informative local descriptors

Authors:
Lucas Paletta;Gerald Fritz;Christin Seifert
Affiliations:
Joanneum Research Forschungsgesellschaft mbH, Wastiangasse, Graz, Austria;Joanneum Research Forschungsgesellschaft mbH, Wastiangasse, Graz, Austria;Joanneum Research Forschungsgesellschaft mbH, Wastiangasse, Graz, Austria
Venue:
ICML '05 Proceedings of the 22nd international conference on Machine learning
Year:
2005

Citing 4
Cited 6

Technical Note: \cal Q-Learning

Machine Learning
A reinforcement learning model of selective visual attention

Proceedings of the fifth international conference on Autonomous agents
Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
Rapid object recognition from discriminative regions of interest

AAAI'04 Proceedings of the 19th national conference on Artifical intelligence

Perception and Developmental Learning of Affordances in Autonomous Robots

KI '07 Proceedings of the 30th annual German conference on Advances in Artificial Intelligence
Learning Object Representations Using Sequential Patterns

AI '08 Proceedings of the 21st Australasian Joint Conference on Artificial Intelligence: Advances in Artificial Intelligence
Closed-loop learning of visual control policies

Journal of Artificial Intelligence Research
Reinforcement learning of predictive features in affordance perception

Proceedings of the 2006 international conference on Towards affordance-based robot control
Learning active fusion of multiple experts' decisions: An attention-based approach

Neural Computation
Context-Aware Semi-Local Feature Detector

ACM Transactions on Intelligent Systems and Technology (TIST)

Quantified Score

Hi-index	0.00

Visualization

Abstract

This work provides a framework for learning sequential attention in real-world visual object recognition, using an architecture of three processing stages. The first stage rejects irrelevant local descriptors based on an information theoretic saliency measure, providing candidates for foci of interest (FOI). The second stage investigates the information in the FOI using a codebook matcher and providing weak object hypotheses. The third stage integrates local information via shifts of attention, resulting in chains of descriptor-action pairs that characterize object discrimination. A Q-learner adapts then from explorative search and evaluative feedback from entropy decreases on the attention sequences, eventually prioritizing shifts that lead to a geometry of descriptor-action scanpaths that is highly discriminative with respect to object recognition. The methodology is successfully evaluated on indoors (COIL-20 database) and outdoors (TSG-20 database) imagery, demonstrating significant impact by learning, outperforming standard local descriptor based methods both in recognition accuracy and processing time.