- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi, guys
The title explain it all. How do we create a recognizable pose/gesture in RealSense?
I've been searching for tutorial for how to make a new gesture in RealSense and no where to find. I know the SDK give us some gesture like v sign, spreadfingers, swipe, etc., but I want to make my own gesture.
The farthest that I've done my research is the hand tracking use the finger coordinates and compare it to the world coordinates to recognize the gesture. Do we have to modify the library or something else? I hope you someone can help me with an instruction, reference, or maybe a sample how to do it. Sorry for my English~
Radit
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Radit,
As far as I am aware, the SDK does not provide a way of adding more gestures.
You can of course read the raw data from the camera, process the data and generate your own events/gestures.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks for your reply, sebastian
Yes, I already think about that method, but I can't find the raw data, or I just simply don't know which one is and how to get it. Can you explain more a little about this raw data from the camera? Thank you in advance
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Take a look at the samples, in particular the one that reads raw data (raw_streams). The source code is available at the SDK installation folder.

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page