Proposal generation for object detection using cascaded ranking SVMs
Object recognition has made great strides recently. However, the best methods, such as those based on kernel-SVMs are highly computationally intensive. The problem of how to accelerate the evaluation process without decreasing accuracy is thus of current interest. In this paper, we deal with this pr...
| Main Authors: | , , |
|---|---|
| Format: | Conference item |
| Language: | English |
| Published: |
IEEE
2011
|
| _version_ | 1817931136807993344 |
|---|---|
| author | Zhang, Z Warrell, J Torr, PHS |
| author_facet | Zhang, Z Warrell, J Torr, PHS |
| author_sort | Zhang, Z |
| collection | OXFORD |
| description | Object recognition has made great strides recently. However, the best methods, such as those based on kernel-SVMs are highly computationally intensive. The problem of how to accelerate the evaluation process without decreasing accuracy is thus of current interest. In this paper, we deal with this problem by using the idea of ranking. We propose a cascaded architecture which using the ranking SVM generates an ordered set of proposals for windows containing object instances. The top ranking windows may then be fed to a more complex detector. Our experiments demonstrate that our approach is robust, achieving higher overlap-recall values using fewer output proposals than the state-of-the-art. Our use of simple gradient features and linear convolution indicates that our method is also faster than the state-of-the-art. |
| first_indexed | 2024-12-09T03:17:14Z |
| format | Conference item |
| id | oxford-uuid:af399b71-c977-4bef-bc60-a0d9d90fa68c |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-12-09T03:17:14Z |
| publishDate | 2011 |
| publisher | IEEE |
| record_format | dspace |
| spelling | oxford-uuid:af399b71-c977-4bef-bc60-a0d9d90fa68c2024-10-22T13:17:20ZProposal generation for object detection using cascaded ranking SVMsConference itemhttp://purl.org/coar/resource_type/c_5794uuid:af399b71-c977-4bef-bc60-a0d9d90fa68cEnglishSymplectic ElementsIEEE2011Zhang, ZWarrell, JTorr, PHSObject recognition has made great strides recently. However, the best methods, such as those based on kernel-SVMs are highly computationally intensive. The problem of how to accelerate the evaluation process without decreasing accuracy is thus of current interest. In this paper, we deal with this problem by using the idea of ranking. We propose a cascaded architecture which using the ranking SVM generates an ordered set of proposals for windows containing object instances. The top ranking windows may then be fed to a more complex detector. Our experiments demonstrate that our approach is robust, achieving higher overlap-recall values using fewer output proposals than the state-of-the-art. Our use of simple gradient features and linear convolution indicates that our method is also faster than the state-of-the-art. |
| spellingShingle | Zhang, Z Warrell, J Torr, PHS Proposal generation for object detection using cascaded ranking SVMs |
| title | Proposal generation for object detection using cascaded ranking SVMs |
| title_full | Proposal generation for object detection using cascaded ranking SVMs |
| title_fullStr | Proposal generation for object detection using cascaded ranking SVMs |
| title_full_unstemmed | Proposal generation for object detection using cascaded ranking SVMs |
| title_short | Proposal generation for object detection using cascaded ranking SVMs |
| title_sort | proposal generation for object detection using cascaded ranking svms |
| work_keys_str_mv | AT zhangz proposalgenerationforobjectdetectionusingcascadedrankingsvms AT warrellj proposalgenerationforobjectdetectionusingcascadedrankingsvms AT torrphs proposalgenerationforobjectdetectionusingcascadedrankingsvms |