Measuring objects in real time using TensorFlow

Hello, wondering if any experts in computer vision or people with experience can help with a project I’ve been working on, meant to measure objects using reference objects in real-time. I have a theory that if you can use a reference object with known size, then you can use the pixel size/ratio of the quarter to be able to measure the objects by their pixel ratio as well so that you can be able to get the relative size

Are the reference object and the target object coplanar?

Yes they would both be laying horizontal on a table or bed or floor or something


Do you think it would be possible if reference was laid on top on the other? like a card of known size, laid on top of a piece of paper I’m trying to measure? both on the table?

Yes it could be possible.

would it be possible to discuss the algorithm more in depth through direct messaging?

We prefer to handle topics that could be useful to other community members.

So you could try to expose your doubts here if they are still TF related.