{"url":"/dataset/crowd11","name":"Crowd 11","full_name":"A Dataset for Fine Grained Crowd Behaviour Analysis","description_markdown":"This dataset defines a total of 11 crowd motion patterns and it is composed of over 6000 video sequences with an average length of 100 frames per sequence.\r\nThis documentation presents how to download and process the Crowd-11 dataset.\r\n\r\nIf you use this dataset, please cite our paper:\r\n```\r\nCamille Dupont, Luis Tobias, and Bertrand Luvison. \"Crowd-11: A Dataset for Fine Grained Crowd Behaviour Analysis.\" In  Computer Vision and Pattern Recognition Workshops (CVPRW), 2017.\r\n```\r\n\r\n\r\nSince this dataset is a composition of web videos and already existing datasets, we ask you to download and accept licence of each source and dataset. \r\nThe construction of the Crowd-11 dataset is composed of two steps:\r\n\r\n# Step 1: Retrieve videos of interest from the web and/or pre-existing datasets\r\n\r\n\r\n## Retrieve the pre-existing datasets of interest\r\n\r\nThe pre-existing datasets are:\r\n\r\n| DATASET NAME       \t| url           \t\t\t\t\t\t\t\t\t \t\t\t\t| $SOURCE_NAME  |\r\n| ------------- \t \t|:-----------------------------------------------------------------:| -------------:|\r\n| UMN      \t\t\t\t| http://mha.cs.umn.edu/proj_events.shtml#crowd \t  \t\t\t\t| umn |\r\n| AGORASET     \t\t\t|https://github.com/piallai/Agoraset | agoraset |\r\n| PETS \t\t\t\t\t| http://www.cvg.reading.ac.uk/PETS2009/a.html#s3     \t\t\t\t| pets |\r\n| HOCKEY FIGHT  \t\t| http://visilab.etsii.uclm.es/personas/oscar/FightDetection/       | hockey |\r\n| MOVIES  \t\t\t\t| http://visilab.etsii.uclm.es/personas/oscar/FightDetection/       | peliculas |\r\n| CUHK  \t\t\t\t| http://www.ee.cuhk.edu.hk/~jshao/CUHKcrowd_files/cuhk_crowd_dataset.htm  | cuhk |\r\n| WWW  \t\t\t\t\t| http://www.ee.cuhk.edu.hk/~jshao/WWWCrowdDataset.html      \t\t| www |\r\n| WORLDEXPO'10 CROWD COUNTING  | http://www.ee.cuhk.edu.hk/~xgwang/expo.html      \t\t\t| shanghai |\r\n| VIOLENT-FLOWS  \t\t| http://www.openu.ac.il/home/hassner/data/violentflows/      \t\t| violent_flow |\r\n\r\n\r\n\r\nThese datasets should be stored in their \"existing_datasets/$SOURCE_NAME/\" folder:  \r\n.  \r\n└── existing_datasets  \r\n    ├── agoraset  \r\n    ├── cuhk  \r\n    ├── hockey  \r\n    ├── peliculas  \r\n    │   ├── fights  \r\n    │   └── noFights  \r\n    ├── pets  \r\n    ├── shanghai  \r\n    ├── umn  \r\n    ├── violent_flow  \r\n    └── www  \r\n\r\n\r\n## Copy the videos of interest from the datasets of interest\r\n\r\nThe list of the videos of interest is in existing_datasets_urls.csv. To extract them into the VOI folder, execute:\r\n```\r\npython existing_datasets_gathering.py\r\n```\r\nThe VOI folder should have the following structure:  \r\n.  \r\n└── VOI  \r\n    ├── agoraset  \r\n    ├── cuhk  \r\n    ├── hockey  \r\n    ├── peliculas  \r\n    ├── pets  \r\n    ├── shanghai  \r\n    ├── umn  \r\n    ├── violent_flow  \r\n    └── www  \r\n\r\n## Download the videos of interest from the web\r\n\r\n\r\nThe web sources are:\r\n\r\n| SOURCE NAME       \t| url           \t\t\t\t\t | $SOURCE_NAME  |\r\n| ------------- \t \t|:----------------------------------:| -------------:|\r\n| YOUTUBE      \t\t\t| https://www.youtube.com/ \t  \t\t | youtube       |\r\n| GETTYIMAGES     \t\t| http://www.gettyimages.fr/ \t\t | gettyimages   |\r\n| POND5 \t\t\t\t| https://www.pond5.com/     \t\t | pond5         |\r\n\r\n\r\n\r\nThe list of the web urls to download is in web_urls.csv.\r\nThe web_urls.csv file's structure is as follows :  \r\n\r\n| $SOURCE NAME       \t| URL           \t    | OUTPUT_NAME  | TS_MULTIPLIER  |\r\n| ------------- \t \t|:---------------------:| ------------:| --------------:|\r\n\r\n\r\nWe do not provide the script to download them, but many tools exist to do it (pytube, urllib, etc...). \r\nNote: a few videos have a ts_multiplier field. These video are in slow motion and the ts_multiplier is provided to speed them up (cf. SETPTS option in avconv).\r\n\r\nThe downloaded videos should be stored in their VOI/$SOURCE_NAME folder, which should now have the following structure:  \r\n.  \r\n└── VOI  \r\n    ├── agoraset  \r\n    ├── cuhk  \r\n    ├── gettyimages  \r\n    ├── hockey  \r\n    ├── peliculas  \r\n    ├── pets  \r\n    ├── pond5  \r\n    ├── shanghai  \r\n    ├── umn  \r\n    ├── violent_flow  \r\n    ├── youtube  \r\n    └── www  \r\n\r\n\r\n\r\n# Step 2: Processing original videos into the Crowd-11 dataset\r\n\r\nOnce the VOI folder is complete, a preprocessing step is required in order to crop and trim the original videos into the Crowd-11 dataset.\r\n\r\nThe preprocessing.csv file's structure is as follows :  \r\n\r\n| Videoname     | Label     | Frame_start  | Frame_end  | Top_left  |  Top_right | Width | Height | $SOURCE_NAME | Scene_number | Crop_number |\r\n| ------------- |:---------:| ------------:| ----------:| ---------:| ----------:| -----:| ------:| ------------:| ------------:| -----------:|\r\n\r\nInstallation:\r\n\r\nYou need to have avconv installed: \r\n```\r\nsudo apt-get install avconv\r\n```\r\nThen, you need to install several python package. A virtualeenv installation is recommended:\r\n```\r\nvirtualenv -p python3 py\r\nsource py/bin/activate\r\npip install sk-video\r\n```\r\n\r\nExecution (in the virtualenv):\r\n```\r\npython script_formating.py\r\n```","description_withheld":null,"homepage":"https://drive.google.com/drive/u/1/folders/1OpWXzczB-6CN-A7n95BkYKIaZimmcvcL","introduced_date":"2017-07-01","introduced_date_note":null,"introduced_by":null,"license":{"name":"Since this dataset is a composition of web videos and already existing datasets, we ask you to download and accept licence of each source and dataset.","url":null},"modalities":[{"name":"Images","url":"/datasets/modality/images"},{"name":"Videos","url":"/datasets/modality/videos"}],"tasks":[{"name":"Video Classification","url":"/task/video-classification","datasets_with_task":"/datasets/task/video-classification"},{"name":"Video Recognition","url":"/task/video-recognition","datasets_with_task":"/datasets/task/video-recognition"}],"languages":[{"name":"English","url":"/datasets/language/english"}],"variants":["Crowd 11"],"data_loaders":[{"repo":"https://github.com/Kaggle/kaggle-api","url":"https://www.kaggle.com/datasets/angeliqueloesch/crowd-11","frameworks":[]}],"num_papers_in_archive":0,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[],"papers_with_a_benchmark_row":[],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":0,"samples_harvested":0,"samples_ran":0,"samples_unverified":0,"pointer_only_for_licence":0,"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."}