Modified centroid triplet loss for person re-identification
Abstract Person Re-identification (ReID) is the process of matching target individuals to their images within different images or videos captured from a variety of angles or cameras. This is a critical task for surveillance applications, in particular, these applications that operate in large enviro...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2023-05-01
|
| Series: | Journal of Big Data |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s40537-023-00753-0 |
| _version_ | 1797817895484391424 |
|---|---|
| author | Alaa Alnissany Yazan Dayoub |
| author_facet | Alaa Alnissany Yazan Dayoub |
| author_sort | Alaa Alnissany |
| collection | DOAJ |
| description | Abstract Person Re-identification (ReID) is the process of matching target individuals to their images within different images or videos captured from a variety of angles or cameras. This is a critical task for surveillance applications, in particular, these applications that operate in large environments such as malls and airports. Recent studies use data-driven approaches to tackle this problem. This work continues on this path by presenting a modification of a previously defined loss, the centroid triplet loss ( CTL). The proposed loss, modified centroid triplet loss (MCTL), emphasizes more on the interclass distance. It is divided into two parts, one penalizes for interclass distance and second penalizes for intraclass distance. Mean Average Precision (mAP) was adopted to validate our approach, two datasets are also used for validation; Market-1501 and DukeMTMC. The results were calculated for first rank of identification and mAP. For dataset Market-1501 dataset, the results were $$98.4\%$$ 98.4 % rank1, $$98.63\%$$ 98.63 % mAP, and $$96.8\%$$ 96.8 % rank1, $$97.3\%$$ 97.3 % mAP on DukeMTMC dataset, the results outweighed those of existing studies in the domain. |
| first_indexed | 2024-03-13T09:01:22Z |
| format | Article |
| id | doaj.art-2ae73adbb131441a87433597660810ee |
| institution | Directory Open Access Journal |
| issn | 2196-1115 |
| language | English |
| last_indexed | 2024-03-13T09:01:22Z |
| publishDate | 2023-05-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Journal of Big Data |
| spelling | doaj.art-2ae73adbb131441a87433597660810ee2023-05-28T11:19:18ZengSpringerOpenJournal of Big Data2196-11152023-05-0110111210.1186/s40537-023-00753-0Modified centroid triplet loss for person re-identificationAlaa Alnissany0Yazan Dayoub1Department of Electronic and Mechanical Systems, Higher Institute for Applied Sciences and TechnologyDepartment of Computer Science, HSE UniversityAbstract Person Re-identification (ReID) is the process of matching target individuals to their images within different images or videos captured from a variety of angles or cameras. This is a critical task for surveillance applications, in particular, these applications that operate in large environments such as malls and airports. Recent studies use data-driven approaches to tackle this problem. This work continues on this path by presenting a modification of a previously defined loss, the centroid triplet loss ( CTL). The proposed loss, modified centroid triplet loss (MCTL), emphasizes more on the interclass distance. It is divided into two parts, one penalizes for interclass distance and second penalizes for intraclass distance. Mean Average Precision (mAP) was adopted to validate our approach, two datasets are also used for validation; Market-1501 and DukeMTMC. The results were calculated for first rank of identification and mAP. For dataset Market-1501 dataset, the results were $$98.4\%$$ 98.4 % rank1, $$98.63\%$$ 98.63 % mAP, and $$96.8\%$$ 96.8 % rank1, $$97.3\%$$ 97.3 % mAP on DukeMTMC dataset, the results outweighed those of existing studies in the domain.https://doi.org/10.1186/s40537-023-00753-0Person ReIDTriplet lossCenter lossInter class distanceCentroid triplet lossDukeMTMC-ReID |
| spellingShingle | Alaa Alnissany Yazan Dayoub Modified centroid triplet loss for person re-identification Journal of Big Data Person ReID Triplet loss Center loss Inter class distance Centroid triplet loss DukeMTMC-ReID |
| title | Modified centroid triplet loss for person re-identification |
| title_full | Modified centroid triplet loss for person re-identification |
| title_fullStr | Modified centroid triplet loss for person re-identification |
| title_full_unstemmed | Modified centroid triplet loss for person re-identification |
| title_short | Modified centroid triplet loss for person re-identification |
| title_sort | modified centroid triplet loss for person re identification |
| topic | Person ReID Triplet loss Center loss Inter class distance Centroid triplet loss DukeMTMC-ReID |
| url | https://doi.org/10.1186/s40537-023-00753-0 |
| work_keys_str_mv | AT alaaalnissany modifiedcentroidtripletlossforpersonreidentification AT yazandayoub modifiedcentroidtripletlossforpersonreidentification |