Nonmyopic View Planning for Active Object Detection
Nikolay A. Atanasov, Bharath Sankaran, Jerome Le Ny, George Pappas, Kostas Daniilidis

Citation
Nikolay A. Atanasov, Bharath Sankaran, Jerome Le Ny, George Pappas, Kostas Daniilidis. "Nonmyopic View Planning for Active Object Detection". IEEE Transactions on Robotics (TRO), September 2013; 12 pages (two-column); 7 figures; 2 tables; Manuscript submitted to the IEEE Transactions on Robotics (TRO).

Abstract
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 views, which balances the amount of energy used to move with the chance of identifying the correct hypothesis. We formulate an active 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 real-world experiments with the PR2 robot. The results suggest that our approach outperforms the widely-used greedy view point selection and provides a significant improvement over static object detection.

Electronic downloads

Citation formats  
  • HTML
    Nikolay A. Atanasov, Bharath Sankaran, Jerome Le Ny, George
    Pappas, Kostas Daniilidis. <a
    href="http://www.terraswarm.org/pubs/125.html"
    >Nonmyopic View Planning for Active Object
    Detection</a>, <i>IEEE Transactions on Robotics
    (TRO)</i>, September 2013; 12 pages (two-column); 7
    figures; 2 tables; Manuscript submitted to the IEEE
    Transactions on Robotics (TRO).
  • Plain text
    Nikolay A. Atanasov, Bharath Sankaran, Jerome Le Ny, George
    Pappas, Kostas Daniilidis. "Nonmyopic View Planning for
    Active Object Detection". <i>IEEE Transactions on
    Robotics (TRO)</i>, September 2013; 12 pages
    (two-column); 7 figures; 2 tables; Manuscript submitted to
    the IEEE Transactions on Robotics (TRO).
  • BibTeX
    @article{AtanasovSankaranLeNyPappasDaniilidis13_NonmyopicViewPlanningForActiveObjectDetection,
        author = {Nikolay A. Atanasov and Bharath Sankaran and
                  Jerome Le Ny and George Pappas and Kostas
                  Daniilidis},
        title = {Nonmyopic View Planning for Active Object Detection},
        journal = {IEEE Transactions on Robotics (TRO)},
        month = {September},
        year = {2013},
        note = {12 pages (two-column); 7 figures; 2 tables;
                  Manuscript submitted to the IEEE Transactions on
                  Robotics (TRO)},
        abstract = {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 views, which balances the amount of
                  energy used to move with the chance of identifying
                  the correct hypothesis. We formulate an active
                  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
                  real-world experiments with the PR2 robot. The
                  results suggest that our approach outperforms the
                  widely-used greedy view point selection and
                  provides a significant improvement over static
                  object detection.},
        URL = {http://terraswarm.org/pubs/125.html}
    }
    

Posted by Nikolay A. Atanasov on 29 Sep 2013.

Notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright.