How to recognize a face from the image of a person using TK1 board?

Robotics & Edge Computing Jetson & Embedded Systems Jetson TK1
1

how to deploy a deep neural network in TK1 board?

Thanks in advance!
Regards,
Saurabh Pahune

2

Hi,

It’s NOT recommended to run deep learning use case on the TK1.
Most frameworks require an 64-bit environment but TK1 only support 32-bit.

You can give the vision-based face recognition algorithm a try.
For example: [url]https://realpython.com/face-detection-in-python-using-a-webcam/[/url]

Thanks.

3

Thanks for an infromation!

Aastall,

do you know how this below features works in TK1 simultaneously or Do I need to change a board for making all below things work parallel?

which framework or process is good to deploy all below features in a en embedded board or any suggestion about an another board?
how will it work?

Features:

1.Body shape recognition: Recognize a person from the shape of his body.
2.Posture recognition: Recognize a person from the posture of his body.
3.Style recognition: Recognize a person from the style of his physical actions (e.g. walking, sitting) and speech (e.g. talking, laughing, crying).

Thanks in advance!

Regards,
Saurabh Pahune

4

Hi,

Sorry for the late reply.

TK1 is not able to do deep learning based inference.
To have a better recognition result, it is recommended to get a more powerful board.

Some example for your reference: NVIDIA AI IOT · GitHub

Thanks.