Human detection with Grid-eye

I am working with Grid-eye sensor from Panasonic and I have some challenges for human detection with Panasonic algorithms and counting them.
1- In a different environments we got different result and accuracy.
2- How to distinguish between human and animals such as dogs?

It would be my pleasure if any body can share experiences with these challenges :innocent:

Hi, I can see this thread is quite old. But Iā€™d like to know if there has been any progress. I am not familiar with the Panasonic sensor. I have seen many similar projects done with Raspberry Pi and Raspberry Pi camera.