Friends Meet Dataset

Friends Meet (FM) Dataset

We present a new annotated dataset that contains groups of people that evolve, appear and disappear spontaneously, and experience split and merge events. This correspond to cocktail party-like situations in social areas where people arrive alone or with other people, move from one group to another, stay still while conversing, etc.  Such a picture was missing in literature, since most of the existing datasets with labelled groups contain wlaking pedestrian with a main flow direction. We manually annotated people and groups on each frame and over time, i.e., detection and tracking of groups and individuals.

 FM dataset is composed by 53 sequences, for a total of 16286 frames. The sequences are partitioned in a synthetic set (28 sequences, 200 frames each) without any complex object representation and dynamics (simple colored blobs), and a real dataset. In the synthetic set, 18 sequences are simple, containing 1-2 events with 4-10 individuals; the other 10 sequences are more challenging, with 10-16 individuals involved in multiple events.

The real set focuses on the outdoor area in IIT where people usually meet during coffee breaks. This area has been recorded and annotated by an expert for one month. The expert reported the events appeared more frequently, building a screenplay where these events are summarized in order to limit the dataset size. Therefore, the screenplay was played by students and employees, resulting in 15 sequences of different length (between 30 sec. to 1.5 minutes), judged by the expert as sufficiently realistic. In total, the sequences contain from 3 to 11 individuals.


Sample images

Friends Meet (FM) Dataset


How to get the dataset

To obtain this dataset, we ask you to complete, sign and return the form below. After that, I will send you the credentials to download it. Note that the dataset is available only for research purposes.

  author = {Bazzani, L. and Murino, V. and Cristani, M.},
  title = {Decentralized Particle Filter for Joint Individual-Group Tracking},
  booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2012},
  month = {June}

Instrunctions are in README.txt. You will require MATLAB to use the data.


Video results




  • L. Bazzani, V. Murino, and M. Cristani
    "Decentralized particle filter for joint individual-group tracking"
    Conference on Computer Vision and Pattern Recognition, (CVPR), 2012 [PDF]