Exploiting feature dynamics for active object recognition
This paper describes a new approach to object recognition for active vision systems that integrates information across multiple observations of an object. The approach exploits the order relationship between successive frames to derive a classifier based on the characteristic motion of local feature...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
Institute of Electrical and Electronics Engineers
2013
|
| Online Access: | http://hdl.handle.net/1721.1/80839 https://orcid.org/0000-0002-4333-7194 https://orcid.org/0000-0002-3315-2372 |
| _version_ | 1811083009148846080 |
|---|---|
| author | Robbel, Philipp Roy, Deb K |
| author2 | Massachusetts Institute of Technology. Media Laboratory |
| author_facet | Massachusetts Institute of Technology. Media Laboratory Robbel, Philipp Roy, Deb K |
| author_sort | Robbel, Philipp |
| collection | MIT |
| description | This paper describes a new approach to object recognition for active vision systems that integrates information across multiple observations of an object. The approach exploits the order relationship between successive frames to derive a classifier based on the characteristic motion of local features across visual sweeps. This motion model reveals structural information about the object that can be exploited for recognition. The main contribution of this paper is a recognition system that extends invariant local features (shape contexts) into the time domain by integration of a motion model. Evaluations on one standardized and one custom collected dataset from the humanoid robot in our laboratory demonstrate that the motion model allows higher-quality hypotheses about object categories quicker than a baseline system that treats object views as unordered streams of images. |
| first_indexed | 2024-09-23T12:17:45Z |
| format | Article |
| id | mit-1721.1/80839 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T12:17:45Z |
| publishDate | 2013 |
| publisher | Institute of Electrical and Electronics Engineers |
| record_format | dspace |
| spelling | mit-1721.1/808392022-09-28T00:55:37Z Exploiting feature dynamics for active object recognition Robbel, Philipp Roy, Deb K Massachusetts Institute of Technology. Media Laboratory Program in Media Arts and Sciences (Massachusetts Institute of Technology) Robbel, Philipp Roy, Deb K. This paper describes a new approach to object recognition for active vision systems that integrates information across multiple observations of an object. The approach exploits the order relationship between successive frames to derive a classifier based on the characteristic motion of local features across visual sweeps. This motion model reveals structural information about the object that can be exploited for recognition. The main contribution of this paper is a recognition system that extends invariant local features (shape contexts) into the time domain by integration of a motion model. Evaluations on one standardized and one custom collected dataset from the humanoid robot in our laboratory demonstrate that the motion model allows higher-quality hypotheses about object categories quicker than a baseline system that treats object views as unordered streams of images. 2013-09-20T16:35:26Z 2013-09-20T16:35:26Z 2010-12 Article http://purl.org/eprint/type/ConferencePaper 978-1-4244-7814-9 9781424478132 1424478138 INSPEC Accession Number: 11805719 http://hdl.handle.net/1721.1/80839 Robbel, Philipp, and Deb Roy. “Exploiting feature dynamics for active object recognition.” In 2010 11th International Conference on Control Automation Robotics & Vision, Singapore, 7-10th December 2010. p.2102-2108. Institute of Electrical and Electronics Engineers, 2010. https://orcid.org/0000-0002-4333-7194 https://orcid.org/0000-0002-3315-2372 en_US http://dx.doi.org/10.1109/ICARCV.2010.5707373 2010 11th International Conference on Control Automation Robotics & Vision Creative Commons Attribution-Noncommercial-Share Alike 3.0 http://creativecommons.org/licenses/by-nc/3.0/ application/pdf Institute of Electrical and Electronics Engineers MIT Web Domain |
| spellingShingle | Robbel, Philipp Roy, Deb K Exploiting feature dynamics for active object recognition |
| title | Exploiting feature dynamics for active object recognition |
| title_full | Exploiting feature dynamics for active object recognition |
| title_fullStr | Exploiting feature dynamics for active object recognition |
| title_full_unstemmed | Exploiting feature dynamics for active object recognition |
| title_short | Exploiting feature dynamics for active object recognition |
| title_sort | exploiting feature dynamics for active object recognition |
| url | http://hdl.handle.net/1721.1/80839 https://orcid.org/0000-0002-4333-7194 https://orcid.org/0000-0002-3315-2372 |
| work_keys_str_mv | AT robbelphilipp exploitingfeaturedynamicsforactiveobjectrecognition AT roydebk exploitingfeaturedynamicsforactiveobjectrecognition |