Revision history [back]

so, you need the fov and the distance, to calculate the size.

and you probably don't have any of it, given arbitrary images as input.

so, you need the fov and the distance, to calculate the size.

and you probably don't have any of it, given arbitrary images as input.

so, there's fov, size and distance, you need the fov and the distance, to know 2 of them, to calculate the size.third.

and you probably don't have any of it, given arbitrary images as input.

(as long as it's like that -- what yu you want is not possible)

so, there's fov, size and distance, you need to know 2 of them, to calculate the third.

(you can get the fov from calibrating your camera, also you can get the distance, if you're using a stereo rig or a depth cam, like kinect)

and you probably don't have any of it, given arbitrary images as input.

(as long as it's like that -- what you want is not possible)

so, there's fov, size and distance, you need to know 2 of them, to calculate the third.

and you probably don't have any of it, given arbitrary images as input.

(you can get the fov from calibrating your camera, also you can get the distance, if you're using a stereo rig or a depth cam, like kinect)

and you probably don't have any of it, given arbitrary images as input.

(as long as it's like that -- what you want is not possible)

so, there's fov, size and distance, you need to know 2 of them, to calculate the third.

and you probably don't have any of it, given arbitrary images as input.

(you can get the fov from calibrating your camera, also you can get the distance, if you're using a stereo rig or a depth cam, like kinect)

(as (again, as long as it's like that long as you use arbitrary images, and have no control about the camera params -- what you want is simply not possible)