Unified crowd segmentation
This paper presents a unified approach to crowd segmentation. A global solution is generated using an Expectation Maximization framework. Initially, a head and shoulder detector is used to nominate an exhaustive set of person locations and these form the person hypotheses. The image is then partitio...
| Main Authors: | , , , , , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
2008
|
| _version_ | 1797087729559797760 |
|---|---|
| author | Tu, P Sebastian, T Doretto, G Krahnstoever, N Rittscher, J Yu, T |
| author_facet | Tu, P Sebastian, T Doretto, G Krahnstoever, N Rittscher, J Yu, T |
| author_sort | Tu, P |
| collection | OXFORD |
| description | This paper presents a unified approach to crowd segmentation. A global solution is generated using an Expectation Maximization framework. Initially, a head and shoulder detector is used to nominate an exhaustive set of person locations and these form the person hypotheses. The image is then partitioned into a grid of small patches which are each assigned to one of the person hypotheses. A key idea of this paper is that while whole body monolithic person detectors can fail due to occlusion, a partial response to such a detector can be used to evaluate the likelihood of a single patch being assigned to a hypothesis. This captures local appearance information without having to learn specific appearance models. The likelihood of a pair of patches being assigned to a person hypothesis is evaluated based on low level image features such as uniform motion fields and color constancy. During the E-step, the single and pairwise likelihoods are used to compute a globally optimal set of assignments of patches to hypotheses. In the M-step, parameters which enforce global consistency of assignments are estimated. This can be viewed as a form of occlusion reasoning. The final assignment of patches to hypotheses constitutes a segmentation of the crowd. The resulting system provides a global solution that does not require background modeling and is robust with respect to clutter and partial occlusion. © 2008 Springer Berlin Heidelberg. |
| first_indexed | 2024-03-07T02:39:49Z |
| format | Journal article |
| id | oxford-uuid:aa0cec1a-d46b-49dc-b6ae-83ebd556dd4a |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T02:39:49Z |
| publishDate | 2008 |
| record_format | dspace |
| spelling | oxford-uuid:aa0cec1a-d46b-49dc-b6ae-83ebd556dd4a2022-03-27T03:12:34ZUnified crowd segmentationJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:aa0cec1a-d46b-49dc-b6ae-83ebd556dd4aEnglishSymplectic Elements at Oxford2008Tu, PSebastian, TDoretto, GKrahnstoever, NRittscher, JYu, TThis paper presents a unified approach to crowd segmentation. A global solution is generated using an Expectation Maximization framework. Initially, a head and shoulder detector is used to nominate an exhaustive set of person locations and these form the person hypotheses. The image is then partitioned into a grid of small patches which are each assigned to one of the person hypotheses. A key idea of this paper is that while whole body monolithic person detectors can fail due to occlusion, a partial response to such a detector can be used to evaluate the likelihood of a single patch being assigned to a hypothesis. This captures local appearance information without having to learn specific appearance models. The likelihood of a pair of patches being assigned to a person hypothesis is evaluated based on low level image features such as uniform motion fields and color constancy. During the E-step, the single and pairwise likelihoods are used to compute a globally optimal set of assignments of patches to hypotheses. In the M-step, parameters which enforce global consistency of assignments are estimated. This can be viewed as a form of occlusion reasoning. The final assignment of patches to hypotheses constitutes a segmentation of the crowd. The resulting system provides a global solution that does not require background modeling and is robust with respect to clutter and partial occlusion. © 2008 Springer Berlin Heidelberg. |
| spellingShingle | Tu, P Sebastian, T Doretto, G Krahnstoever, N Rittscher, J Yu, T Unified crowd segmentation |
| title | Unified crowd segmentation |
| title_full | Unified crowd segmentation |
| title_fullStr | Unified crowd segmentation |
| title_full_unstemmed | Unified crowd segmentation |
| title_short | Unified crowd segmentation |
| title_sort | unified crowd segmentation |
| work_keys_str_mv | AT tup unifiedcrowdsegmentation AT sebastiant unifiedcrowdsegmentation AT dorettog unifiedcrowdsegmentation AT krahnstoevern unifiedcrowdsegmentation AT rittscherj unifiedcrowdsegmentation AT yut unifiedcrowdsegmentation |