genisysai / vision2018 Goto Github PK

Currently the process when a user is detected is as follows:

• User is matched against known database of images or marked as intruder
• TASS communicates with iotJumpWay and stores the user ID / name, activity and classification / location etc in the iotJumpWay database.
• TASS devices are able to track the user if there is a camera unit in the room
• Applications can then connect to that database via the API to access data that they are authorised to access

A coming update will swap this activity to using a local database provided and integrated with a future release / push. Scope of this move will include:

• Removal of all identifying information with exception to location on local network and classification (Static Values: Known/Unknown) that is sent to iotJumpWay.
• Changing storage location of user activity data to local MySQL database opposed to iotJumpWay global database.

This move will mean all data that could be used to track trends never leaves your home network but can be utilised by your custom applications.

Related code:

https://github.com/GeniSysAI/Vision/blob/master/Local/LocalStreamer.py

Thank you for this useful implementation of real-time face recognition on the edge!

I would like to ask if there are any plans for introducing compatibility of this program with the Intel Neural Compute Stick 2 and the OpenVINO toolkit, instead of the Movidius NCS and NCSDK?

Thanks and cheers!