Sequential optimization for efficient high-quality object proposal generation
We are motivated by the need for a generic object proposal generation algorithm which achieves good balance between object detection recall, proposal localization quality and computational efficiency. We propose a novel object proposal algorithm, BING++, which inherits the virtue of good computation...
| Auteurs principaux: | , , , , , , |
|---|---|
| Format: | Journal article |
| Publié: |
IEEE
2017
|
| _version_ | 1826300336822812672 |
|---|---|
| author | Zhang, Z Liu, Y Chen, X Zhu, Y Cheng, M Saligrama, V Torr, P |
| author_facet | Zhang, Z Liu, Y Chen, X Zhu, Y Cheng, M Saligrama, V Torr, P |
| author_sort | Zhang, Z |
| collection | OXFORD |
| description | We are motivated by the need for a generic object proposal generation algorithm which achieves good balance between object detection recall, proposal localization quality and computational efficiency. We propose a novel object proposal algorithm, BING++, which inherits the virtue of good computational efficiency of BING [1] but significantly improves its proposal localization quality. At high level we formulate the problem of object proposal generation from a novel probabilistic perspective, based on which our BING++ manages to improve the localization quality by employing edges and segments to estimate object boundaries and update the proposals sequentially. We propose learning the parameters efficiently by searching for approximate solutions in a quantized parameter space for complexity reduction. We demonstrate the generalization of BING++ with the same fixed parameters across different object classes and datasets. Empirically our BING++ can run at half speed of BING on CPU, but significantly improve the localization quality by 18.5% and 16.7% on both VOC2007 and Microhsoft COCO datasets, respectively. Compared with other state-of-the-art approaches, BING++ can achieve comparable performance, but run significantly faster. |
| first_indexed | 2024-03-07T05:15:36Z |
| format | Journal article |
| id | oxford-uuid:dd14c48e-23c3-43e7-befb-9052f8f74f50 |
| institution | University of Oxford |
| last_indexed | 2024-03-07T05:15:36Z |
| publishDate | 2017 |
| publisher | IEEE |
| record_format | dspace |
| spelling | oxford-uuid:dd14c48e-23c3-43e7-befb-9052f8f74f502022-03-27T09:22:23ZSequential optimization for efficient high-quality object proposal generationJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:dd14c48e-23c3-43e7-befb-9052f8f74f50Symplectic Elements at OxfordIEEE2017Zhang, ZLiu, YChen, XZhu, YCheng, MSaligrama, VTorr, PWe are motivated by the need for a generic object proposal generation algorithm which achieves good balance between object detection recall, proposal localization quality and computational efficiency. We propose a novel object proposal algorithm, BING++, which inherits the virtue of good computational efficiency of BING [1] but significantly improves its proposal localization quality. At high level we formulate the problem of object proposal generation from a novel probabilistic perspective, based on which our BING++ manages to improve the localization quality by employing edges and segments to estimate object boundaries and update the proposals sequentially. We propose learning the parameters efficiently by searching for approximate solutions in a quantized parameter space for complexity reduction. We demonstrate the generalization of BING++ with the same fixed parameters across different object classes and datasets. Empirically our BING++ can run at half speed of BING on CPU, but significantly improve the localization quality by 18.5% and 16.7% on both VOC2007 and Microhsoft COCO datasets, respectively. Compared with other state-of-the-art approaches, BING++ can achieve comparable performance, but run significantly faster. |
| spellingShingle | Zhang, Z Liu, Y Chen, X Zhu, Y Cheng, M Saligrama, V Torr, P Sequential optimization for efficient high-quality object proposal generation |
| title | Sequential optimization for efficient high-quality object proposal generation |
| title_full | Sequential optimization for efficient high-quality object proposal generation |
| title_fullStr | Sequential optimization for efficient high-quality object proposal generation |
| title_full_unstemmed | Sequential optimization for efficient high-quality object proposal generation |
| title_short | Sequential optimization for efficient high-quality object proposal generation |
| title_sort | sequential optimization for efficient high quality object proposal generation |
| work_keys_str_mv | AT zhangz sequentialoptimizationforefficienthighqualityobjectproposalgeneration AT liuy sequentialoptimizationforefficienthighqualityobjectproposalgeneration AT chenx sequentialoptimizationforefficienthighqualityobjectproposalgeneration AT zhuy sequentialoptimizationforefficienthighqualityobjectproposalgeneration AT chengm sequentialoptimizationforefficienthighqualityobjectproposalgeneration AT saligramav sequentialoptimizationforefficienthighqualityobjectproposalgeneration AT torrp sequentialoptimizationforefficienthighqualityobjectproposalgeneration |