Datasets:

bassatbassat
/

GroOT-MOT17

video_id stringlengths 14 14	video stringlengths 25 25	captions listlengths 1 1	caption_types listlengths 1 1	subset stringclasses 1 value	split stringclasses 1 value	sequence stringclasses 7 values	track_id int64 1 170	num_frames int64 8 300
MOT17-02_t0002	videos/MOT17-02_t0002.mp4	[ "man walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	2	56
MOT17-02_t0003	videos/MOT17-02_t0003.mp4	[ "woman walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	3	277
MOT17-02_t0008	videos/MOT17-02_t0008.mp4	[ "woman walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	8	36
MOT17-02_t0009	videos/MOT17-02_t0009.mp4	[ "a person sitting on a bench" ]	[ "action" ]	groot_mot17	test	MOT17-02	9	256
MOT17-02_t0010	videos/MOT17-02_t0010.mp4	[ "a woman standing on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	10	277
MOT17-02_t0014	videos/MOT17-02_t0014.mp4	[ "woman walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	14	225
MOT17-02_t0015	videos/MOT17-02_t0015.mp4	[ "man walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	15	213
MOT17-02_t0016	videos/MOT17-02_t0016.mp4	[ "a woman pulling luggage" ]	[ "action" ]	groot_mot17	test	MOT17-02	16	57
MOT17-02_t0017	videos/MOT17-02_t0017.mp4	[ "a person standing on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	17	236
MOT17-02_t0018	videos/MOT17-02_t0018.mp4	[ "person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	18	203
MOT17-02_t0019	videos/MOT17-02_t0019.mp4	[ "woman walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	19	300
MOT17-02_t0020	videos/MOT17-02_t0020.mp4	[ "woman walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	20	300
MOT17-02_t0021	videos/MOT17-02_t0021.mp4	[ "woman walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	21	73
MOT17-02_t0022	videos/MOT17-02_t0022.mp4	[ "person walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	22	10
MOT17-02_t0023	videos/MOT17-02_t0023.mp4	[ "woman walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	23	300
MOT17-02_t0026	videos/MOT17-02_t0026.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	26	300
MOT17-02_t0027	videos/MOT17-02_t0027.mp4	[ "person walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	27	281
MOT17-02_t0028	videos/MOT17-02_t0028.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	28	168
MOT17-02_t0029	videos/MOT17-02_t0029.mp4	[ "person standing on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	29	300
MOT17-02_t0031	videos/MOT17-02_t0031.mp4	[ "person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	31	151
MOT17-02_t0032	videos/MOT17-02_t0032.mp4	[ "a man walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	32	270
MOT17-02_t0033	videos/MOT17-02_t0033.mp4	[ "person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	33	219
MOT17-02_t0034	videos/MOT17-02_t0034.mp4	[ "woman walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	34	240
MOT17-02_t0035	videos/MOT17-02_t0035.mp4	[ "man walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	35	221
MOT17-02_t0036	videos/MOT17-02_t0036.mp4	[ "person standing on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	36	269
MOT17-02_t0037	videos/MOT17-02_t0037.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	37	137
MOT17-02_t0038	videos/MOT17-02_t0038.mp4	[ "person pacing on pavement" ]	[ "action" ]	groot_mot17	test	MOT17-02	38	243
MOT17-02_t0039	videos/MOT17-02_t0039.mp4	[ "person pacing on pavement" ]	[ "action" ]	groot_mot17	test	MOT17-02	39	261
MOT17-02_t0040	videos/MOT17-02_t0040.mp4	[ "person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	40	134
MOT17-02_t0041	videos/MOT17-02_t0041.mp4	[ "person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	41	122
MOT17-02_t0042	videos/MOT17-02_t0042.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	42	104
MOT17-02_t0043	videos/MOT17-02_t0043.mp4	[ "person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	43	236
MOT17-02_t0044	videos/MOT17-02_t0044.mp4	[ "a person walking" ]	[ "action" ]	groot_mot17	test	MOT17-02	44	178
MOT17-02_t0048	videos/MOT17-02_t0048.mp4	[ "a person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	48	153
MOT17-02_t0049	videos/MOT17-02_t0049.mp4	[ "person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	49	144
MOT17-02_t0054	videos/MOT17-02_t0054.mp4	[ "person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	54	198
MOT17-02_t0055	videos/MOT17-02_t0055.mp4	[ "person walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	55	37
MOT17-02_t0056	videos/MOT17-02_t0056.mp4	[ "a person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	56	96
MOT17-02_t0057	videos/MOT17-02_t0057.mp4	[ "a person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	57	155
MOT17-02_t0058	videos/MOT17-02_t0058.mp4	[ "woman walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	58	167
MOT17-02_t0059	videos/MOT17-02_t0059.mp4	[ "person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	59	125
MOT17-02_t0060	videos/MOT17-02_t0060.mp4	[ "person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	60	108
MOT17-02_t0061	videos/MOT17-02_t0061.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	61	40
MOT17-02_t0062	videos/MOT17-02_t0062.mp4	[ "person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-02	62	145
MOT17-02_t0067	videos/MOT17-02_t0067.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	67	11
MOT17-02_t0074	videos/MOT17-02_t0074.mp4	[ "a person on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	74	68
MOT17-02_t0079	videos/MOT17-02_t0079.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-02	79	9
MOT17-04_t0001	videos/MOT17-04_t0001.mp4	[ "man standing and waiting" ]	[ "action" ]	groot_mot17	test	MOT17-04	1	300
MOT17-04_t0002	videos/MOT17-04_t0002.mp4	[ "man walking on a sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	2	300
MOT17-04_t0003	videos/MOT17-04_t0003.mp4	[ "man standing and waiting on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	3	300
MOT17-04_t0004	videos/MOT17-04_t0004.mp4	[ "man talking to his cell phone" ]	[ "action" ]	groot_mot17	test	MOT17-04	4	300
MOT17-04_t0005	videos/MOT17-04_t0005.mp4	[ "person walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	5	300
MOT17-04_t0006	videos/MOT17-04_t0006.mp4	[ "man using cell phone" ]	[ "action" ]	groot_mot17	test	MOT17-04	6	300
MOT17-04_t0060	videos/MOT17-04_t0060.mp4	[ "person standing on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	60	300
MOT17-04_t0061	videos/MOT17-04_t0061.mp4	[ "man walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	61	80
MOT17-04_t0062	videos/MOT17-04_t0062.mp4	[ "person standing on the pavement" ]	[ "action" ]	groot_mot17	test	MOT17-04	62	300
MOT17-04_t0063	videos/MOT17-04_t0063.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	63	300
MOT17-04_t0065	videos/MOT17-04_t0065.mp4	[ "a person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	65	300
MOT17-04_t0066	videos/MOT17-04_t0066.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	66	300
MOT17-04_t0067	videos/MOT17-04_t0067.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	67	300
MOT17-04_t0068	videos/MOT17-04_t0068.mp4	[ "person standing on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	68	300
MOT17-04_t0069	videos/MOT17-04_t0069.mp4	[ "person standing on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	69	300
MOT17-04_t0070	videos/MOT17-04_t0070.mp4	[ "a man walking and carrying a bag" ]	[ "action" ]	groot_mot17	test	MOT17-04	70	300
MOT17-04_t0071	videos/MOT17-04_t0071.mp4	[ "a person walking on a street" ]	[ "action" ]	groot_mot17	test	MOT17-04	71	300
MOT17-04_t0072	videos/MOT17-04_t0072.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	72	53
MOT17-04_t0073	videos/MOT17-04_t0073.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	73	51
MOT17-04_t0074	videos/MOT17-04_t0074.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	74	300
MOT17-04_t0075	videos/MOT17-04_t0075.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	75	300
MOT17-04_t0076	videos/MOT17-04_t0076.mp4	[ "a person standing on sidewalk and taking photos" ]	[ "action" ]	groot_mot17	test	MOT17-04	76	300
MOT17-04_t0077	videos/MOT17-04_t0077.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	77	16
MOT17-04_t0078	videos/MOT17-04_t0078.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	78	300
MOT17-04_t0080	videos/MOT17-04_t0080.mp4	[ "a person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	80	300
MOT17-04_t0082	videos/MOT17-04_t0082.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	82	300
MOT17-04_t0083	videos/MOT17-04_t0083.mp4	[ "a man walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	83	300
MOT17-04_t0084	videos/MOT17-04_t0084.mp4	[ "person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	84	266
MOT17-04_t0085	videos/MOT17-04_t0085.mp4	[ "person walking on the pavement" ]	[ "action" ]	groot_mot17	test	MOT17-04	85	276
MOT17-04_t0086	videos/MOT17-04_t0086.mp4	[ "a person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	86	300
MOT17-04_t0087	videos/MOT17-04_t0087.mp4	[ "person standing on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	87	300
MOT17-04_t0088	videos/MOT17-04_t0088.mp4	[ "a person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	88	300
MOT17-04_t0089	videos/MOT17-04_t0089.mp4	[ "a person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	89	300
MOT17-04_t0090	videos/MOT17-04_t0090.mp4	[ "a person standing on a sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	90	300
MOT17-04_t0091	videos/MOT17-04_t0091.mp4	[ "a person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	91	300
MOT17-04_t0092	videos/MOT17-04_t0092.mp4	[ "a person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	92	300
MOT17-04_t0093	videos/MOT17-04_t0093.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	93	300
MOT17-04_t0094	videos/MOT17-04_t0094.mp4	[ "a person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	94	300
MOT17-04_t0095	videos/MOT17-04_t0095.mp4	[ "a person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	95	216
MOT17-04_t0096	videos/MOT17-04_t0096.mp4	[ "a person standing on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	96	38
MOT17-04_t0098	videos/MOT17-04_t0098.mp4	[ "person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	98	300
MOT17-04_t0099	videos/MOT17-04_t0099.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	99	300
MOT17-04_t0100	videos/MOT17-04_t0100.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	100	300
MOT17-04_t0101	videos/MOT17-04_t0101.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	101	300
MOT17-04_t0102	videos/MOT17-04_t0102.mp4	[ "a person walking on the sidewalk and carrying a bag" ]	[ "action" ]	groot_mot17	test	MOT17-04	102	200
MOT17-04_t0103	videos/MOT17-04_t0103.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	103	202
MOT17-04_t0104	videos/MOT17-04_t0104.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	104	56
MOT17-04_t0105	videos/MOT17-04_t0105.mp4	[ "man walking on sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	105	300
MOT17-04_t0106	videos/MOT17-04_t0106.mp4	[ "man walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	106	300
MOT17-04_t0107	videos/MOT17-04_t0107.mp4	[ "person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	107	300
MOT17-04_t0108	videos/MOT17-04_t0108.mp4	[ "a person walking on the sidewalk" ]	[ "action" ]	groot_mot17	test	MOT17-04	108	234
MOT17-04_t0109	videos/MOT17-04_t0109.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	109	33
MOT17-04_t0110	videos/MOT17-04_t0110.mp4	[ "a person walking on the street" ]	[ "action" ]	groot_mot17	test	MOT17-04	110	18

End of preview. Expand in Data Studio

GroOT-MOT17 tracklets

A text-to-tracklet retrieval benchmark for MOT17 pedestrians. Each captioned MOT17 track is cropped into its own short mp4. A text query must retrieve the matching track from a gallery of 454 tracks. The layout is the same as bassatbassat/TVPReid, so the same evaluation code runs on both.

This is an unofficial derivative. The videos come from MOT17 and the captions come from GroOT. The credit for both belongs to their authors (see below). What this repository adds is listed under Our contribution.

Configs

All configs share the same 454-track gallery. Only the queries differ.

Config	Queries	Unique query texts	What a query is
`appearance`	344	299	GroOT appearance caption, as released
`action`	454	112	GroOT action caption, as released
`all`	798	411	both of the above
`combined`	344	304	one refined sentence per track describing appearance and action (ours)

Tracks without a caption of the requested type stay in the gallery as distractors.

Read the action numbers with care. GroOT action captions are mostly generic. The most common one, "a person walking on the street", is shared by 78 tracks. Only 112 of 454 action captions are unique. An action query often matches several tracks equally well, yet exactly one counts as correct. The combined config was made to reduce this ambiguity.

Original work (credit)

Captions: GroOT, from Type-to-Track: Retrieve Any Object via Prompt-based Tracking by Pha Nguyen, Kha Gia Quach, Kris Kitani, and Khoa Luu (NeurIPS 2023). GroOT gives every MOT17 box captions = [appearance, action]. We use mot17_train_coco.json from uark-cviu/Type-to-Track. The authors mark the MOT17 test captions as sub-optimal, so we do not use them.
Videos and boxes: MOT17 / MOT16 by Anton Milan, Laura Leal-Taixé, Ian Reid, Stefan Roth, and Konrad Schindler (motchallenge.net).

The appearance and action captions in this repository are the GroOT captions unchanged, apart from whitespace trimming and dropping empty or None values.

Our contribution

Tracklet videos. Each GroOT track is cropped from the MOT17 FRCNN frames into one mp4, so the track itself can be retrieved instead of individual boxes.
Retrieval splits. The appearance, action, all, and combined query sets share one gallery, in the TVPReid format.
Refined combined captions (combined). For the 344 tracks that have both captions, the two GroOT captions were merged into one fluent sentence about a single subject. It works in two passes:
- A rule-based merge (combine_captions in the build code) joins the two captions and keeps the more specific subject ("woman" over "person").
- An LLM assistant (Cursor agent, Auto model routing) then reviewed all 344 merged captions one by one. It fixed grammar and articles, removed repeated phrases such as "carrying a bag, walking down the street carrying two bags", and resolved two gender conflicts between a track's appearance and action captions by looking at the frames (MOT17-05 tracks 54 and 59). No new visual attributes were added. Every refined caption uses only facts already present in the two GroOT captions.
The refined captions are versioned in groot_mot17_combined.json.
Build code, so the dataset can be regenerated from the original sources: scripts/build_groot_mot17.py.

Construction

Only the FRCNN copy of each sequence is used, because DPM and SDP share the same frames.
Boxes with ignore=1, height < 50px, or visibility < 0.25 are dropped.
Tracks need at least 8 boxes and at least one caption. Long tracks are subsampled uniformly to at most 300 boxes.
Each box is grown to the track's median aspect ratio, so the extra width is real scene context rather than black padding. It is then resized to height 256 and written at the sequence frame rate.

Layout

videos/<MOT17-XX>_t<track>.mp4
metadata/all-test.jsonl
metadata/appearance-test.jsonl
metadata/action-test.jsonl
metadata/combined-test.jsonl

Each row has these fields: video_id, video, captions, caption_types, subset, split, sequence, track_id, num_frames.

Usage

# pip install "shawaf-vlm @ git+https://github.com/BASSAT-BASSAT/Benchmarking-Video-Image-language-models-for-Tracklet-retrieval-.git"
from shawaf_vlm.data import load_groot

splits = load_groot("combined")  # downloads the metadata and mp4s on first use
print(len(splits.query), len(splits.gallery))

Standard protocol: each caption is a query, and its own track is the only positive. Report Rank-1/5/10, mAP, and median rank over the whole gallery.

Intended use and limitations

Zero-shot evaluation only. There is no train split, and GroOT's MOT17 test captions are not included.
Captions describe a single pedestrian in a crowded street, mall, or hallway. Many people wear similar clothes, so some queries are still ambiguous even in combined (for example "a person in a black jacket walking on the sidewalk").
The combined captions are machine-refined. Wording can differ from what a human annotator would write, but the content is limited to the source captions.

License

GroOT annotations and MOT17 videos are released under CC BY-NC-SA 3.0. This dataset, including the refined captions, keeps that license: non-commercial use, with attribution, share-alike.

Citation

If you use this dataset, cite GroOT and MOT17. If you use the tracklet splits or the combined captions, please also cite this repository.

@article{nguyen2023type,
  title   = {Type-to-Track: Retrieve Any Object via Prompt-based Tracking},
  author  = {Nguyen, Pha and Quach, Kha Gia and Kitani, Kris and Luu, Khoa},
  journal = {Advances in Neural Information Processing Systems},
  year    = {2023}
}
@article{milan2016mot16,
  title   = {MOT16: A Benchmark for Multi-Object Tracking},
  author  = {Milan, Anton and Leal-Taix{\'e}, Laura and Reid, Ian and Roth, Stefan and Schindler, Konrad},
  journal = {arXiv:1603.00831},
  year    = {2016}
}
@misc{shawaf_groot_mot17,
  title        = {GroOT-MOT17 tracklets: text-to-tracklet retrieval splits with refined combined captions},
  howpublished = {\url{https://huggingface.co/datasets/bassatbassat/GroOT-MOT17}},
  note         = {Derived from GroOT (Nguyen et al., 2023) and MOT17. Code: https://github.com/BASSAT-BASSAT/Benchmarking-Video-Image-language-models-for-Tracklet-retrieval-},
  year         = {2026}
}

Downloads last month: 45

Paper for bassatbassat/GroOT-MOT17

MOT16: A Benchmark for Multi-Object Tracking

Paper • 1603.00831 • Published Mar 2, 2016