{"url":"/dataset/mars","name":"MARS","full_name":"Motion Analysis and Re-identification Set","description_markdown":"**MARS** (**Motion Analysis and Re-identification Set**) is a large scale video based person reidentification dataset, an extension of the Market-1501 dataset. It has been collected from six near-synchronized cameras. It consists of 1,261 different pedestrians, who are captured by at least 2 cameras. The variations in poses, colors and illuminations of pedestrians, as well as the poor image quality, make it very difficult to yield high matching accuracy. Moreover, the dataset contains 3,248 distractors in order to make it more realistic. Deformable Part Model and GMMCP tracker were used to automatically generate the tracklets (mostly 25-50 frames long).\r\n\r\nSource: [Multi-Target Tracking in Multiple Non-Overlapping Cameras using Constrained Dominant Sets](https://arxiv.org/abs/1706.06196)","description_withheld":null,"homepage":"http://zheng-lab.cecs.anu.edu.au/Project/project_mars.html","introduced_date":"2016-01-01","introduced_date_note":null,"introduced_by":{"paper":null,"title":"MARS: A Video Benchmark for Large-Scale Person Re-Identification","first_author":null,"url":"https://doi.org/10.1007/978-3-319-46466-4_52"},"license":{"name":"Unknown","url":null},"modalities":[{"name":"Videos","url":"/datasets/modality/videos"}],"tasks":[{"name":"Person Re-Identification","url":"/task/person-re-identification","datasets_with_task":"/datasets/task/person-re-identification"},{"name":"Unsupervised Person Re-Identification","url":"/task/unsupervised-person-re-identification","datasets_with_task":"/datasets/task/unsupervised-person-re-identification"},{"name":"Video-Based Person Re-Identification","url":"/task/video-based-person-re-identification","datasets_with_task":"/datasets/task/video-based-person-re-identification"}],"languages":[],"variants":["MARS"],"data_loaders":[{"repo":"https://github.com/activeloopai/Hub","url":"https://docs.activeloop.ai/datasets/mars-dataset","frameworks":["tf","pytorch"]}],"num_papers_in_archive":181,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/person-re-identification-on-mars","task":"Person Re-Identification","dataset_variant":"MARS","rows":21,"metrics":["mAP","Rank-1","Rank-5","Rank-10","Rank-20"],"first_row_in_archive_order":{"model":"B-BOT + OSM + CL Centers* (Re-rank)","paper":"/paper/video-person-re-id-fantastic-techniques-and","metrics":{"mAP":"88.5"},"code_links":[{"title":"ppriyank/Video-Person-Re-ID-Fantastic-Techniques-and-Where-to-Find-Them","url":"https://github.com/ppriyank/Video-Person-Re-ID-Fantastic-Techniques-and-Where-to-Find-Them"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/unsupervised-person-re-identification-on-mars","task":"Unsupervised Person Re-Identification","dataset_variant":"MARS","rows":1,"metrics":["mAP","rank-1"],"first_row_in_archive_order":{"model":"AuxUSLReID","paper":"/paper/a-high-accuracy-unsupervised-person-re","metrics":{"mAP":"72.4","rank-1":"80.9"},"code_links":[{"title":"tenghehan/AuxUSLReID","url":"https://github.com/tenghehan/AuxUSLReID"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/a-high-accuracy-unsupervised-person-re","title":"A High-Accuracy Unsupervised Person Re-identification Method Using Auxiliary Information Mined from Datasets","date":"2022-05-06","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/multi-direction-and-multi-scale-pyramid-in-1","title":"Multi-direction and Multi-scale Pyramid in Transformer for Video-based Pedestrian Retrieval","date":"2022-02-12","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/video-based-person-re-identification-with-2","title":"Video-based Person Re-identification with Spatial and Temporal Memory Networks","date":"2021-08-20","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/spatio-temporal-representation-factorization","title":"Spatio-Temporal Representation Factorization for Video-based Person Re-Identification","date":"2021-07-25","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/learning-multi-granular-hypergraphs-for-video-1","title":"Learning Multi-Granular Hypergraphs for Video-Based Person Re-Identification","date":"2021-04-30","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":1,"samples_ran":1,"samples_unverified":0,"pointer_only_for_licence":1,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/cloth-changing-person-re-identification-from","title":"Cloth-Changing Person Re-identification from A Single Image with Gait Prediction and Regularization","date":"2021-03-29","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/dense-interaction-learning-for-video-based","title":"Dense Interaction Learning for Video-based Person Re-identification","date":"2021-03-16","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/pyramid-spatial-temporal-aggregation-for","title":"Pyramid Spatial-Temporal Aggregation for Video-Based Person Re-Identification","date":"2021-01-01","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/fine-grained-re-identification","title":"Fine-Grained Re-Identification","date":"2020-11-26","rows_on_this_dataset":1,"code_links":2,"syntology":null},{"paper":"/paper/robust-re-identification-by-multiple-views","title":"Robust Re-Identification by Multiple Views Knowledge Distillation","date":"2020-07-08","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/video-person-re-id-fantastic-techniques-and","title":"Video Person Re-ID: Fantastic Techniques and Where to Find Them","date":"2019-11-21","rows_on_this_dataset":3,"code_links":1,"syntology":null},{"paper":"/paper/adaptive-graph-representation-learning-for","title":"Adaptive Graph Representation Learning for Video Person Re-identification","date":"2019-09-05","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/temporal-knowledge-propagation-for-image-to","title":"Temporal Knowledge Propagation for Image-to-Video Person Re-identification","date":"2019-08-11","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/spatially-and-temporally-efficient-non-local","title":"Spatially and Temporally Efficient Non-local Attention Network for Video-based Person Re-Identification","date":"2019-08-05","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/unsupervised-tracklet-person-re","title":"Unsupervised Tracklet Person Re-Identification","date":"2019-03-01","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/a-pose-sensitive-embedding-for-person-re","title":"A Pose-Sensitive Embedding for Person Re-Identification with Expanded Cross Neighborhood Re-Ranking","date":"2017-11-28","rows_on_this_dataset":1,"code_links":2,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":3,"samples_ran":1,"samples_unverified":2,"pointer_only_for_licence":3,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/in-defense-of-the-triplet-loss-for-person-re","title":"In Defense of the Triplet Loss for Person Re-Identification","date":"2017-03-22","rows_on_this_dataset":4,"code_links":31,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":46,"samples_ran":6,"samples_unverified":40,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":3,"samples_harvested":50,"samples_ran":8,"samples_unverified":42,"pointer_only_for_licence":4,"papers_with_no_sample_that_ran":0,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}