IEEE SPS Video and Image Processing Cup (VIP-Cup)

Activity recognition from body cameras

VIP-Cup

Organised in conjunction with the IEEE International Conference on Image Processing 2019, Taipei, Taiwan

The increasing availability of wearable cameras enables the collection of first-person videos (FPV) for the recognition of activities at home, in the workplace and during sport activities. FPV activity recognition has important applications, which include assisted living, activity tracking and life-logging. The main challenges of FPV activity recognition are the presence of outlier motions (for example due to other people captured by the camera), motion blur, illumination changes and self-occlusions, which are particularly important with chest-mounted cameras.

The 2019 VIP-Cup challenge focuses on FPV from a chest-mounted camera and on the privacy-aware recognition of activities, which include generic activities, such as walking, person-to-person interactions, such as chatting and handshaking, and person-to-object interactions, such as using a computer or a whiteboard. As videos captured by body cameras may leak private or sensitive information about individuals, the evaluation of the IEEE VIP-Cup challenge entries will include privacy enhancing solutions jointly with the recognition performance.

A dataset of activities from several subjects is provided with the annotation for training and validation (http://www.eecs.qmul.ac.uk/~andrea/fpvo). The evaluation will be performed based on test datasets that will be provided closer to the submission deadline.

A set of MATLAB scripts are provided in GitHub to get started with the tasks: https://github.com/girmaw/VIPCUP.

The call for participants can be found [here].


Participate

Step 1: get familiar with the full set of instructions [here]

Step 2: join the VIP-Cup forum in Piazza, for any Questions and Answers:

  • Go to https://piazza.com
  • Go to the "Sign Up" tab (top-left corner) and select "StudentsGet Stared"
  • In the "Search Schools:", type "IEEE SPS" and select the pop-up option
  • Choose "Summer 2019" as "Selected Term"
  • In "Class 1:" type and select ”VIPCUP2019: IEEE Video and Image Processing (VIP) Cup2019”
  • Insert as "Class Access Code" the string "sps009", select ”Join as: Student” and click ”Join the Class”
  • Complete the "Please enter your email address" fields and click "Submit Email"
  • Check your inbox for your confirmation email and activate your piazza account (via link or validation code)
  • Proceed with setting up your piazza account

Step 3: register to submit your work https://www2.securecms.com/VIPCup/VIPRegistration.asp.


Schedule

30 April 2019: Participation guidelines and initial dataset released
5 May 2019: Additional Training Dataset released
15 June 2019 30 June 2019 : Deadline for result submission
15 July 2019: Finalists (best three teams) announced
30 July 2019: Test Dataset released
22 September 2019: Competition on a new Test Dataset at ICIP 2019

The three finalist teams will participate to the second phase of the competition that will be held at ICIP 2019. Trip and expenses will be covered.


Eligibility

We invite teams satisfying the following eligibility criteria to participate in the VIP-Cup. Each team must be composed of:

  • One faculty member (the Supervisor) - postdocs and research associates are not considered as faculty members;
  • At most one graduate student (the Tutor) - a graduate student is a student having earned at least a 4-year University degree at the time of submission;
  • At least three but no more than ten undergraduate students (the Team Members) - an undergraduate student is a student without a 4-year degree
    • all undergraduate students are eligible to participate (i.e. third and fourth year undergraduate students are eligible)
    • students who are in their fourth year of a 5-year program that will culminate in a Master's degree (and do not hold a Bachelors degree) are eligible to participate as regular undergraduate Team Members;
  • At least three undergraduate members of a finalist team must be either IEEE Signal Processing Society (SPS) members or SPS student members.
The VIP-Cup is a competition for undergraduate students and therefore Master's students, regardless of the duration of their Bachelor's degree, cannot participate as regular Team Members (one graduate student can however be the Tutor for a team).


Team prizes

Winnner: $5,000

1st runner-up: $2,500

2nd runner-up: $1,500


Organisers

Girmaw Abebe Tadesse, University of Oxford (girmaw.abebe@eng.ox.ac.uk)
Oliver Bent, University of Oxford
Kommy Woldemariam, IBM Research Africa
Andrea Cavallaro, Queen Mary University of London

Supported by

IEEE Signal Processing Society (SPS)
Computational Health Informatics Group, Oxford University
IBM Research Africa
Centre for Intelligent Sensing, Queen Mary University of London