Active information acquisition can be used to optimize the trajectory of a depth camera in 3-D space in order to improve the performance of object classification and pose estimation. We can formulate a stochastic optimal control problem in which the objective is to minimize the probability of misclassification subject to constraints from the camera observation and motion model. Our work proposes an exact planning algorithm based on dynamic programming that obtains the optimal camera control policy but scales poorly with the size of the state and measurement spaces. To provide scalability, we developed an approximation algorithm based on Monte Carlo tree search with a rollout policy that exploits the structure of the probability of error objective function. Our experiments suggest that active approaches for camera view planning provide significant improvements over static object recognition. Also, an advantage of non-greedy planning is that it allows high-confidence object recognition with an adaptive decision threshold that depends on the observations received online.
This work proposes an active approach for part-based object detection, which optimizes the order of part filter evaluations and the time at which to stop and make a prediction. Statistics, describing the part responses, are learned from training data and are used to formalize part scheduling as an offline optimal control problem. Dynamic programming is applied to obtain a policy, which balances the number of part evaluations with the classification accuracy. During inference, the policy is used as a look-up table to choose the part order and the stopping time based on the observed filter responses. The method is faster than cascade detection with deformable part models (which does not optimize the part order) with negligible loss in accuracy when evaluated on the PASCAL VOC 2007 and 2010 datasets.
@article{Atanasov_ActiveObjectRecognition_TRO14,
author = {N. Atanasov and B. Sankaran and J. Le Ny and G. Pappas and K. Daniilidis},
title = {Nonmyopic View Planning for Active Object Classification and Pose Estimation},
journal = {IEEE Trans. on Robotics (TRO)},
year = {2014},
volume = {30},
number = {5},
pages = {1078-1090},
doi = {http://www.doi.org/10.1109/TRO.2014.2320795}
}
@inproceedings{Atanasov_Sankaran_ActiveObjectRecognition_ICRA13,
author = {N. Atanasov and B. Sankaran and J. Le Ny and T. Koletschka and G. Pappas and K. Daniilidis},
title = {Hypothesis Testing Framework for Active Object Detection},
booktitle = {IEEE Int. Conf. on Robotics and Automation (ICRA)},
year = {2013},
pages = {4216-4222},
doi = {http://www.doi.org/10.1109/ICRA.2013.6631173}
}
@inproceedings{Zhu_ADPM_ECCV14,
author = {M. Zhu and N. Atanasov and G. Pappas and K. Daniilidis},
title = {Active Deformable Part Models Inference},
booktitle = {European Conf. on Computer Vision (ECCV)},
year = {2014},
volume = {8695},
pages = {281-296},
doi = {http://www.doi.org/10.1007/978-3-319-10584-0_19}
}
@inproceedings{Lauri_MonteCarloTreeSearch_ICRA15_Workshop,
author = {M. Lauri and N. Atanasov and G. Pappas and R. Ritala},
title = {Active Object Recognition via Monte Carlo Tree Search},
booktitle = {Workshop on Beyond Geometric Constraints at ICRA},
year = {2015}
}