Plan-view Trajectory Estimation with Dense Stereo Background Models
In a known environment, objects may be tracked in multiple views using a set of back-ground models. Stereo-based models can be illumination-invariant, but often have undefined values which inevitably lead to foreground classification errors. We derive dense stereo models for object tracking using lo...
| Main Authors: | , , , |
|---|---|
| Language: | en_US |
| Published: |
2004
|
| Online Access: | http://hdl.handle.net/1721.1/6075 |
| _version_ | 1826192631655301120 |
|---|---|
| author | Darrell, T. Demirdjian, D. Checka, N. Felzenswalb, P. |
| author_facet | Darrell, T. Demirdjian, D. Checka, N. Felzenswalb, P. |
| author_sort | Darrell, T. |
| collection | MIT |
| description | In a known environment, objects may be tracked in multiple views using a set of back-ground models. Stereo-based models can be illumination-invariant, but often have undefined values which inevitably lead to foreground classification errors. We derive dense stereo models for object tracking using long-term, extended dynamic-range imagery, and by detecting and interpolating uniform but unoccluded planar regions. Foreground points are detected quickly in new images using pruned disparity search. We adopt a 'late-segmentation' strategy, using an integrated plan-view density representation. Foreground points are segmented into object regions only when a trajectory is finally estimated, using a dynamic programming-based method. Object entry and exit are optimally determined and are not restricted to special spatial zones. |
| first_indexed | 2024-09-23T09:26:20Z |
| id | mit-1721.1/6075 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T09:26:20Z |
| publishDate | 2004 |
| record_format | dspace |
| spelling | mit-1721.1/60752019-04-12T08:28:59Z Plan-view Trajectory Estimation with Dense Stereo Background Models Darrell, T. Demirdjian, D. Checka, N. Felzenswalb, P. In a known environment, objects may be tracked in multiple views using a set of back-ground models. Stereo-based models can be illumination-invariant, but often have undefined values which inevitably lead to foreground classification errors. We derive dense stereo models for object tracking using long-term, extended dynamic-range imagery, and by detecting and interpolating uniform but unoccluded planar regions. Foreground points are detected quickly in new images using pruned disparity search. We adopt a 'late-segmentation' strategy, using an integrated plan-view density representation. Foreground points are segmented into object regions only when a trajectory is finally estimated, using a dynamic programming-based method. Object entry and exit are optimally determined and are not restricted to special spatial zones. 2004-10-04T14:37:37Z 2004-10-04T14:37:37Z 2001-02-01 AIM-2001-001 http://hdl.handle.net/1721.1/6075 en_US AIM-2001-001 5522496 bytes 672260 bytes application/postscript application/pdf application/postscript application/pdf |
| spellingShingle | Darrell, T. Demirdjian, D. Checka, N. Felzenswalb, P. Plan-view Trajectory Estimation with Dense Stereo Background Models |
| title | Plan-view Trajectory Estimation with Dense Stereo Background Models |
| title_full | Plan-view Trajectory Estimation with Dense Stereo Background Models |
| title_fullStr | Plan-view Trajectory Estimation with Dense Stereo Background Models |
| title_full_unstemmed | Plan-view Trajectory Estimation with Dense Stereo Background Models |
| title_short | Plan-view Trajectory Estimation with Dense Stereo Background Models |
| title_sort | plan view trajectory estimation with dense stereo background models |
| url | http://hdl.handle.net/1721.1/6075 |
| work_keys_str_mv | AT darrellt planviewtrajectoryestimationwithdensestereobackgroundmodels AT demirdjiand planviewtrajectoryestimationwithdensestereobackgroundmodels AT checkan planviewtrajectoryestimationwithdensestereobackgroundmodels AT felzenswalbp planviewtrajectoryestimationwithdensestereobackgroundmodels |