I'm currently researching how to create a Hazard Perception system using the Movidius NCS. My manager has specified that he wants me to make it work using TensorFlow trained models.
(Thanks to scarcity of information pertaining to this, neither of us are really sure what's what) - I've been trying to work out whether or not the NCS can run TF models , and so far I've found conflicting answers.
In my use case I'll be needing to process live feed from multiple cameras at about 15FPS (Obviously I'm going to have to work this out in due time)…
Would anyone know if there are TensorFlow neural networks that are no only capable of such speed, but are also supported by the Movidius NCS?
If not, useful links to learning resources for any of the following would be greatly appreciated:
-Object detection on the Movidius NCS
-Training custom models (Preferably with TensorFlow)
-Any alternative systems for Object detection on the Movidius NCS
Thanks in advance!