How is the distance threshold determined?

[davidmurray]

Hello,

In the code, I noticed that 250 is used as a threshold to classify people as being too close ("red"). However, this is a distance in pixels, not meters. How was it determined then?

Also, isn't that ignoring the problem of perspective? (if two people are close to the camera then their pixel distance will be large, but if they are far away from the camera than their pixel distance will be small).

Thanks,
David

[paul-pias]

Hi,
Thank you for reaching out! I didn't clearly understand your first question. Here, The object detector giving me (x,y,w,h) where x & y are the position of the object in the image. So I implemented simple euclidean distance between two objects. The threshold is up to you to determine, I used which seemed convenient to me.

And for your second question, yes it is ignoring the problem of perspective as the depth information is missing from the 2D image. I didn't calculate the distance in pixels so the performance solely depends on how the object detector is performing while finding the saliency of an object in a given frame.

I hope that answers your questions.

[SamiurRahman1]

Hello, i am having a little problem understanding how we know that the distance between two detected persons is 1.5 or 2 meters? How do we determine/proximate the distance in meters from the euclidean distance between two objects?

Thanks