Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313408215> ?p ?o ?g. }
- W4313408215 endingPage "280" @default.
- W4313408215 startingPage "267" @default.
- W4313408215 abstract "Unifying object detection and re-identification (ReID) into a single network enables faster multi-object tracking (MOT), while this multi-task setting poses challenges for training. In this work, we dissect the joint training of detection and ReID from two dimensions: label assignment and loss function. We find previous works generally overlook them and directly borrow the practices from object detection, inevitably causing inferior performance. Specifically, we identify a qualified label assignment for MOT should: 1) have the assignment cost aware of ReID cost, not just detection cost; 2) provide sufficient positive samples for robust feature learning while avoiding ambiguous positives (i.e., the positives shared by different ground-truth objects). To achieve the above goals, we first propose Identity-aware Label Assignment, which jointly considers the assignment cost of detection and ReID to select positive samples for each instance without ambiguities. Moreover, we advance a novel Discriminative Focal Loss that integrates ReID predictions with Focal Loss to focus the training on the discriminative samples. Finally, we upgrade the strong baseline FairMOT with our techniques and achieve up to 7.0 MOTA / 54.1% IDs improvements on MOT16/17/20 benchmarks under favorable inference speed, which verifies our tailored label assignment and loss function for MOT are superior to those inherited from object detection." @default.
- W4313408215 created "2023-01-06" @default.
- W4313408215 creator A5004941473 @default.
- W4313408215 creator A5049471074 @default.
- W4313408215 creator A5051366196 @default.
- W4313408215 creator A5053473760 @default.
- W4313408215 creator A5061811574 @default.
- W4313408215 date "2023-01-01" @default.
- W4313408215 modified "2023-10-03" @default.
- W4313408215 title "A Closer Look at the Joint Training of Object Detection and Re-Identification in Multi-Object Tracking" @default.
- W4313408215 cites W1861492603 @default.
- W4313408215 cites W2107775979 @default.
- W4313408215 cites W2124781496 @default.
- W4313408215 cites W2150066425 @default.
- W4313408215 cites W2150298366 @default.
- W4313408215 cites W2194775991 @default.
- W4313408215 cites W2222512263 @default.
- W4313408215 cites W2511791013 @default.
- W4313408215 cites W2594507094 @default.
- W4313408215 cites W2603203130 @default.
- W4313408215 cites W2739491435 @default.
- W4313408215 cites W2768024897 @default.
- W4313408215 cites W2805447362 @default.
- W4313408215 cites W2894176037 @default.
- W4313408215 cites W2953920664 @default.
- W4313408215 cites W2962766617 @default.
- W4313408215 cites W2963150697 @default.
- W4313408215 cites W2963323244 @default.
- W4313408215 cites W2963351448 @default.
- W4313408215 cites W2963574614 @default.
- W4313408215 cites W2963677766 @default.
- W4313408215 cites W2963901085 @default.
- W4313408215 cites W2964121718 @default.
- W4313408215 cites W2981393651 @default.
- W4313408215 cites W2998792609 @default.
- W4313408215 cites W3028052296 @default.
- W4313408215 cites W3034334609 @default.
- W4313408215 cites W3035314413 @default.
- W4313408215 cites W3035442500 @default.
- W4313408215 cites W3035727180 @default.
- W4313408215 cites W3047363187 @default.
- W4313408215 cites W3095753995 @default.
- W4313408215 cites W3096068180 @default.
- W4313408215 cites W3096609285 @default.
- W4313408215 cites W3099206234 @default.
- W4313408215 cites W3101708136 @default.
- W4313408215 cites W3106763294 @default.
- W4313408215 cites W3110700482 @default.
- W4313408215 cites W3119686997 @default.
- W4313408215 cites W3165926952 @default.
- W4313408215 cites W3167949052 @default.
- W4313408215 cites W3171162369 @default.
- W4313408215 cites W3174690341 @default.
- W4313408215 cites W3175891031 @default.
- W4313408215 cites W3176403636 @default.
- W4313408215 cites W3190647944 @default.
- W4313408215 cites W3207452968 @default.
- W4313408215 cites W3212622989 @default.
- W4313408215 cites W4200316912 @default.
- W4313408215 cites W4313117614 @default.
- W4313408215 doi "https://doi.org/10.1109/tip.2022.3227814" @default.
- W4313408215 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37015359" @default.
- W4313408215 hasPublicationYear "2023" @default.
- W4313408215 type Work @default.
- W4313408215 citedByCount "3" @default.
- W4313408215 countsByYear W43134082152023 @default.
- W4313408215 crossrefType "journal-article" @default.
- W4313408215 hasAuthorship W4313408215A5004941473 @default.
- W4313408215 hasAuthorship W4313408215A5049471074 @default.
- W4313408215 hasAuthorship W4313408215A5051366196 @default.
- W4313408215 hasAuthorship W4313408215A5053473760 @default.
- W4313408215 hasAuthorship W4313408215A5061811574 @default.
- W4313408215 hasConcept C116834253 @default.
- W4313408215 hasConcept C119857082 @default.
- W4313408215 hasConcept C120665830 @default.
- W4313408215 hasConcept C121332964 @default.
- W4313408215 hasConcept C127413603 @default.
- W4313408215 hasConcept C138885662 @default.
- W4313408215 hasConcept C146849305 @default.
- W4313408215 hasConcept C153180895 @default.
- W4313408215 hasConcept C154945302 @default.
- W4313408215 hasConcept C192209626 @default.
- W4313408215 hasConcept C201995342 @default.
- W4313408215 hasConcept C202474056 @default.
- W4313408215 hasConcept C2776151529 @default.
- W4313408215 hasConcept C2776214188 @default.
- W4313408215 hasConcept C2776401178 @default.
- W4313408215 hasConcept C2780451532 @default.
- W4313408215 hasConcept C2781238097 @default.
- W4313408215 hasConcept C31972630 @default.
- W4313408215 hasConcept C41008148 @default.
- W4313408215 hasConcept C41895202 @default.
- W4313408215 hasConcept C59822182 @default.
- W4313408215 hasConcept C64869954 @default.
- W4313408215 hasConcept C86803240 @default.
- W4313408215 hasConcept C97931131 @default.
- W4313408215 hasConceptScore W4313408215C116834253 @default.
- W4313408215 hasConceptScore W4313408215C119857082 @default.