Japanese Page

How to Participate HASC Challenge 2012

Please your Registration:Hub.hasc.jp

HASC Mailing list

You can register in HASC mailing list by following page.
  Google Groups (hasc-jp)
(The mailing list is an invitation system.)

Participation Requirement

To participate HASC Challenge 2012, you need to submit your activity data. After submitting your data, you can get HASC2012corpus which was collected on HASC Challenge 2011.

Challenge Contents

In HASC Challenge 2012, The following challenges are targeted.

  • Challenge 0:Activity data submission. (Please submit one or both of the following contents.)
    • 6 types of activity data (stay, walk, jog, skip, stair-up, stair-down)
    • Real world activity data (Movement between landmarks)

    About data format, please refer to following page Data Format. And, there are sample activity data on Download Page. Acceleration data must contain in your activity data. If you can, please submit another types of data such as gyroscope-data and GPS data.

  • Challenge 1:Activity Recognition (Target: 6 types of activity such as stay, walk, jog, skip, stair-up and stair-down)
  • Challenge 2:Activity Analysis (Presumption of following information)
    • 2-1:The number of steps (Target: walk, jog, stair-up and stair-down)
    • 2-2:Walking-pace(cm) (Target: walk)
    • 2-3:Height of one stair step(cm) (Target: stair-up and stair-down)
  • Challenge 3:Proposal-style
    Participant sets up the challenge content.

About Input/Output Data Format

In Challenge 1-3, input data format is "time(sec) + activity data(csv-format)". Output data format is defined as follows.

  • Challenge 1:Start/End time(sec) and Activity(label data format)
  • Challenge 2:Presumption of the number of steps, walking-pace(cm) or height of one stair step(cm) (meta data format)
  • Challenge 3:Participant sets up the output data format

About Evaluation

The participant receives data for the evaluation , and submits the evaluation result. Or, the participant submits the algorithm, and we evaluates it by the sponsor side.

About Learning Data for Algorithm

You can use your own activity data for your algorithm.

Participation Advantage

You can use the collected data freely.

  • The collected data is not open to the public at once.
Back to Top