Data Description
The dataset was recorded in the Smart Life Care Unit of the Kyushu Institute of Technology in Japan. A motion capture system from Motion Analysis Company is used for this experiment. It has 29 body markers but in the challenge, we are opening only 13 body marker data. 4 subjects (men) in their 20’s and 30’s participated in this experiment. Each of the subjects is instructed to put three types of food in the bento box.
Here they were instructed to perform 5 different scenarios:
- Normal
- Forgot to put ingredients
- Failed to put ingredients
- Turn over bento-box and
- Fix/rearranging ingredients
Actions are done in two different patterns inward and outward. Participants were asked to repeat each work 5 times. 20 infrared cameras are used to track the markers. The three-dimensional position of each marker is recorded. The markers may be labeled incorrectly in some cases due to the complex setting.
- Length of each activity segment: 50~70 s.
- Position of the markers:
- Frequency: 100Hz
- Preprocessing: No preprocessing method is applied to this data
- Training data format:
Activity Name | Label Number in Training Data |
---|---|
Normal(inward) | 1 |
normal(outward) | 2 |
Forgot to put ingredients (inward) | 3 |
Forgot to put ingredients (outward) | 4 |
Failed to put ingredients(inward) | 5 |
Failed to put ingredients(inward) | 6 |
Turn over bento-box (inward) | 7 |
Turn over bento-box (outward) | 8 |
Fix/rearranging ingredients (inward) | 9 |
Fix/rearranging ingredients (outward) | 10 |
- Testng data format: same as training data.