Re-Ranking via metric fusion for object retrieval and person re-identification
This work studies the unsupervised re-ranking procedure for object retrieval and person re-identification with a specific concentration on an ensemble of multiple metrics (or similarities). While the re-ranking step is involved by running a diffusion process on the underlying data manifolds, the fus...
| Hauptverfasser: | , , , |
|---|---|
| Format: | Conference item |
| Veröffentlicht: |
IEEE
2020
|
| _version_ | 1826271508147732480 |
|---|---|
| author | Bai, S Tang, P Torr, P Latecki, L |
| author_facet | Bai, S Tang, P Torr, P Latecki, L |
| author_sort | Bai, S |
| collection | OXFORD |
| description | This work studies the unsupervised re-ranking procedure for object retrieval and person re-identification with a specific concentration on an ensemble of multiple metrics (or similarities). While the re-ranking step is involved by running a diffusion process on the underlying data manifolds, the fusion step can leverage the complementarity of multiple metrics. We give a comprehensive summary of existing fusion with diffusion strategies, and systematically analyze their pros and cons. Based on the analysis, we propose a unified yet robust algorithm which inherits their advantages and discards their disadvantages. Hence, we call it Unified Ensemble Diffusion (UED). More interestingly, we derive that the inherited properties indeed stem from a theoretical framework, where the relevant works can be elegantly summarized as special cases of UED by imposing additional constraints on the objective function and varying the solver of similarity propagation. Extensive experiments with 3D shape retrieval, image retrieval and person re-identification demonstrate that the proposed framework outperforms the state of the arts, and at the same time suggest that re-ranking via metric fusion is a promising tool to further improve the retrieval performance of existing algorithms. |
| first_indexed | 2024-03-06T21:57:45Z |
| format | Conference item |
| id | oxford-uuid:4d81dd3b-fa8d-48c3-91c4-97768f89ffa8 |
| institution | University of Oxford |
| last_indexed | 2024-03-06T21:57:45Z |
| publishDate | 2020 |
| publisher | IEEE |
| record_format | dspace |
| spelling | oxford-uuid:4d81dd3b-fa8d-48c3-91c4-97768f89ffa82022-03-26T15:55:53ZRe-Ranking via metric fusion for object retrieval and person re-identificationConference itemhttp://purl.org/coar/resource_type/c_5794uuid:4d81dd3b-fa8d-48c3-91c4-97768f89ffa8Symplectic Elements at OxfordIEEE2020Bai, STang, PTorr, PLatecki, LThis work studies the unsupervised re-ranking procedure for object retrieval and person re-identification with a specific concentration on an ensemble of multiple metrics (or similarities). While the re-ranking step is involved by running a diffusion process on the underlying data manifolds, the fusion step can leverage the complementarity of multiple metrics. We give a comprehensive summary of existing fusion with diffusion strategies, and systematically analyze their pros and cons. Based on the analysis, we propose a unified yet robust algorithm which inherits their advantages and discards their disadvantages. Hence, we call it Unified Ensemble Diffusion (UED). More interestingly, we derive that the inherited properties indeed stem from a theoretical framework, where the relevant works can be elegantly summarized as special cases of UED by imposing additional constraints on the objective function and varying the solver of similarity propagation. Extensive experiments with 3D shape retrieval, image retrieval and person re-identification demonstrate that the proposed framework outperforms the state of the arts, and at the same time suggest that re-ranking via metric fusion is a promising tool to further improve the retrieval performance of existing algorithms. |
| spellingShingle | Bai, S Tang, P Torr, P Latecki, L Re-Ranking via metric fusion for object retrieval and person re-identification |
| title | Re-Ranking via metric fusion for object retrieval and person re-identification |
| title_full | Re-Ranking via metric fusion for object retrieval and person re-identification |
| title_fullStr | Re-Ranking via metric fusion for object retrieval and person re-identification |
| title_full_unstemmed | Re-Ranking via metric fusion for object retrieval and person re-identification |
| title_short | Re-Ranking via metric fusion for object retrieval and person re-identification |
| title_sort | re ranking via metric fusion for object retrieval and person re identification |
| work_keys_str_mv | AT bais rerankingviametricfusionforobjectretrievalandpersonreidentification AT tangp rerankingviametricfusionforobjectretrievalandpersonreidentification AT torrp rerankingviametricfusionforobjectretrievalandpersonreidentification AT lateckil rerankingviametricfusionforobjectretrievalandpersonreidentification |