Control recognition bounds for visual learning and exploration

Karasev, Vasiliy; Chiuso, Alessandro; Soatto, Stefano

doi:10.1109/ita.2013.6502995

Cited by 12 publications

(15 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Work addressing the multi-view case [16], [17] has shown that merging data from different views leads to improved results, but also highlights the importance of viewpoint selection. From a theoretical perspective, [18] provides a mathematical description of the tradeoff between recognition and planning, which serves as a justification for active viewpoint selection as we explore here.…”

Section: Related Workmentioning

confidence: 99%

Viewpoint Evaluation for Online 3-D Active Object Classification

Patten

Zillich

Fitch

et al. 2016

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

We present an end-to-end method for active object classification in cluttered scenes from RGB-D data. Our algorithms predict the quality of future viewpoints in the form of entropy using both class and pose. Occlusions are explicitly modelled in predicting the visible regions of objects, which modulates the corresponding discriminatory value of a given view. We implement a one-step greedy planner and demonstrate our method online using a mobile robot. We also analyse the performance of our method compared to similar strategies in simulated execution using the Willow Garage dataset. Results show that our active method usefully reduces the number of views required to accurately classify objects in clutter as compared to traditional passive perception.

show abstract

Section: Related Workmentioning

confidence: 99%

Viewpoint Evaluation for Online 3-D Active Object Classification

Patten

Zillich

Fitch

et al. 2016

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

show abstract

“…The cost function is called the value of information [22, ch. 15] or the expected Bayesian risk [23]. Unfortunately, Krause [22,Prop.…”

Section: Related Workmentioning

confidence: 99%

“…Similar to our approach, the problem is encoded by a POMDP, but instead of an approximate nonmyopic policy, the authors resort to a myopic approach to reduce the differential entropy in the pose and class distributions. Karasev et al [23] plan the path of a mobile sensor for visual search of an object in an otherwise known and static scene. The authors hypothesize about the object's pose and apply greedy maximization of the conditional entropy of the next measurement.…”

Section: Related Workmentioning

confidence: 99%

Nonmyopic View Planning for Active Object Classification and Pose Estimation

Atanasov

Sankaran

et al. 2014

IEEE Trans. Robot.

View full text Add to dashboard Cite

One of the central problems in computer vision is the detection of semantically important objects and the estimation of their pose. Most of the work in object detection has been based on single image processing, and its performance is limited by occlusions and ambiguity in appearance and geometry. This paper proposes an active approach to object detection in which the point of view of a mobile depth camera is controlled. When an initial static detection phase identifies an object of interest, several hypotheses are made about its class and orientation. Then, a sequence of views, which balances the amount of energy used to move the sensor with the chance of identifying the correct hypothesis, is planned. We formulate an active hypothesis testing problem, which includes sensor mobility, and solve it using a point-based approximate partially observable Markov decision process algorithm. The validity of our approach is verified through simulation and realworld experiments with the PR2 robot. The results suggest that the approach outperforms the widely used greedy viewpoint selection and provides a significant improvement over static object detection.

show abstract

“…Karasev et al [20] plan the path of a mobile sensor for visual search of an object in an otherwise known and static scene. The problem statement is different from ours but the optimization is surprisingly similar.…”

Section: Related Workmentioning

confidence: 99%

Hypothesis testing framework for active object detection

Atanasov

Sankaran

et al. 2013

2013 IEEE International Conference on Robotics and Automation

View full text Add to dashboard Cite

One of the central problems in computer vision is the detection of semantically important objects and the estimation of their pose. Most of the work in object detection has been based on single image processing and its performance is limited by occlusions and ambiguity in appearance and geometry. This paper proposes an active approach to object detection by controlling the point of view of a mobile depth camera. When an initial static detection phase identifies an object of interest, several hypotheses are made about its class and orientation. The sensor then plans a sequence of viewpoints, which balances the amount of energy used to move with the chance of identifying the correct hypothesis. We formulate an active M-ary hypothesis testing problem, which includes sensor mobility, and solve it using a point-based approximate POMDP algorithm. The validity of our approach is verified through simulation and experiments with real scenes captured by a kinect sensor. The results suggest a significant improvement over static object detection.

show abstract

Control recognition bounds for visual learning and exploration

Cited by 12 publications

References 23 publications

Viewpoint Evaluation for Online 3-D Active Object Classification

Viewpoint Evaluation for Online 3-D Active Object Classification

Nonmyopic View Planning for Active Object Classification and Pose Estimation

Hypothesis testing framework for active object detection

Contact Info

Product

Resources

About