DeepView'21 @ AVSS

Challenges



Introduction

Detailed participant instructions can be accessed here.

Participants can compete in one or more of the following two challenges:

Challenge Track 1: Multi-Camera Multi-Object Counting


Participating teams will count pedestrians from multiple camera scenes. For example, teams will perform pedestrian counting separately for each camera and person re-identification for identifying overlapped ids. This helps traffic surveillance administrators understand the complexity and flows, which can be used to expand better surveillance application plans. To maximize the practical value of the outcome from this track, both object detection and the computational execution efficiency will contribute to the final score for each participating team.The team with the lowest RMSE score in the submitted teams will be announced the winner of this track.

Challenge Track 2: Abnormal Event Detection in Elevator Scenes


Participating teams will submit various abnormal events detected in the test data, including invasion, fighting and shouting. The team with the highest average precision in abnormal event detection in the submitted teams will be announced the winner of this track.

Important Dates

  • 2021.10.20 Testing Results Submission Deadline

  • 2021.10.25 Fact sheets and code/executable submission deadline


Related Links

  • Benchmark and Data: Please contact brightyoun@gist.ac.kr to access the dataset. Before mailing, please complete and attach your application form given here.

  • Results Submission

    • For Challenge Track 1: [Done]

    • For Challenge Track 2: [Done]

  • Paper Submission: [Done]


Submissions

In addition to the testing results, participants are encouraged to submit a paper to DeepView workshop illustrating the challenge methodology and results. All submissions will be handled electronically via the CMT system. By submitting a paper, the authors agree to the policies stipulated on this website. Our challenge calls for papers of two categories: Contributed Papers, and Extended Abstracts. All submissions should be in PDF format. All accepted submissions across the two categories will be invited to give poster presentations. From those accepted, a select few will be invited to give spotlight talks.

Contributed Papers follow a double-blind review process and all the accepted papers will be published as part of the "AVSS Proceedings" and should, therefore, follow the same guideline as the main conference. Paper submission guidelines of AVSS can be accessed here. Extended Abstracts follows single-blind review, and can be up to 4 pages in length (references excluded). This option is meant to provide a non-archival submission option for previously published work, or work that is intended to be published at a future venue. Supplementary appendices are allowed but will be read at the discretion of the reviewers.

Participants are required to submit a supplementary material in .zip format via brightyoun@gist.ac.kr, including the source code, method explanation, and detailed team information etc. Details will appear soon.


Evaluation

  • For each of the two challenge tasks, a different data set will be provided as a set of videos or images. Associated numeric video IDs for each track are obtained by sorting track videos (or the name of the folders in which they are stored) in alphanumeric order, with numbering starting at 1. All the pixel coordinates are 0-based for all tracks.


  • Submission Format for Track 1

    • To be ranked on the public leaderboard of dataset on Track 1, one text file should be submitted to the online evaluation system containing, on each line, details of one counted pedestrian, in the following format (values are comma(',')-delimited):

〈video_id〉 〈frame_id〉 〈movement_id〉 〈pedestrian_class_id〉

where:

  • 〈video_id〉 is the video numeric identifier, starting with 1. It represents the position of the video in the list of all track videos, sorted in alphanumeric order.

  • 〈frame_id〉 represents the frame count for the current frame in the current video, starting with 1.

  • 〈movement_id〉 denotes the the movement numeric identifier, starting with 1. It represents the position of the movement in the list of the MOIs defined in the corresponding instruction document of that video.

  • 〈pedestrian_class_id〉 is the pedestrian classic numeric identifier. Only two values are accepted {1, 2} where 1 stands for “car” and 2 represents “truck”.

  • Online execution: Please submit your result as a csv file (e.g. "track1.csv") through Kaggle.


  • Submission Format for Track 2

    • To be ranked on the public leader board of dataset on Track 2, one text file should be submitted to the online evaluation system containing, on each line, details of one counted pedestrian, in the following format (values are comma(',')-delimited):


〈video_id〉 〈action_category〉

where:

  • 〈video_id〉 is the video numeric identifier, starting with 1. It represents the position of the video in the list of all track videos, sorted in alphanumeric order.

  • 〈action_category〉 can be divided into 6 types: 1) assault, 2) fainting, 3) panic, 4) illegal flyer, 5) escape and 6) normal.

  • Online execution: Please submit your result as a csv file (e.g. "track2.csv") through Kaggle.