Training Objects for Realsense Cameras

[SSchm9]

Apologies if this is a repost, but realized this may be more of a topic for this forum vs. "the Deep Learning Training Tool forum"... a tad confusing....

Working with the following: 

- ZR300 Camera
- Ubuntu 16.04
- QT - CMake
- librealsense SDK used.

---------------------------------------

Does anyone have recommendations to implementing an object training process for the Realsense environment. I am just beginning to review the Deep Learning Tool Kit, but see no documentation for deploying models that work with Realsense cameras.... or create my own classifier file for object recognition.

My ultimate goal is to track multiple people and my trained objects...

Thanks

[Anna_B_Intel · Accepted solution]

Hi Stephen, 

Deep Learning Deployment Toolkit is a software for inference of already trained NN model. For training purpose you may look at Intel optimized frameworks Caffe and Tensorflow. 

Best wishes, 

Anna

[SSchm9]

Anne,

That was exactly what I was looking for. Thanks. 

I was able to install the "Training Kit" - Windows/ Ubuntu on AWS and run a couple of test models, but now about 3-4 hours in, the site is unresponsive. No changes in AWS and performance of the web base toolkit has come to a screeching halt...

Any thoughts?

Thanks

[SSchm9]

Anne,

Looks like I am having AWS EC2 issues. You have satisfied my original post.

Thanks