simonwsw
commented
7 years ago I Update the readme and add a few lines about the dataset arrangement.
In short, the 2 in file name like 0_2_* means the gesture collection session ID. It could be a session collected by different people or different session of a same people. Different session could have different number of files.
- 11 (gestures) 25 (instances) 10 (users) for cross user evaluation: session 2 (25), 3 (25), 5 (25), 6 (25), 8 (25), 9 (25), 10 (25), 11 (25), 12 (25), 13 (25).
- 11 (gestures) (50 (instances) 4 (sessions) + 25 (instances) * 2 (sessions)) for single user cross session evaluation: session 0 (50), 1 (50), 4 (50), 7 (50), 13 (25), 14 (25).
Gesture ID 11 is background class. I suggest you collect your own background class for proper training.
Hello, I want to know overall organization of your dataset.
For example, there are 50 files with a filename starting with 00 and 25 files with a filename starting with 02.
Why they are different?
Also, I couldn't understand why there are 275 files with a filename starting with 1115.
What does the gesture ID 11 mean in your dataset?
I thought you considered 11 different gestures so that the gesture ID value is between 0 and 10.
Thank you.