Questions about neural networks?

Question

Neural networks

Hello, questions about photogrammetry with a neural network.

I have an idea.

1. Take a photo and use a neural network to find objects in the image.
   
2. Determine the size of objects and compare with real sizes.
   
3. Using the dimensions and characteristics of the camera, find how far they are from the camera.
   
4. The result will not be exact, but this is what I need.
   

Will it work and if so, how to implement it.

The second question: What are the self-learning algorithms other than the "perceptron". Similar to this ORG algorithm .

Thank you in advance for your response.

Answer 1

The question is not clear: "Will it work". You gave a list of three tasks and one of your reflections. Each of these tasks is quite solvable and discussed even on this forum many times. Well, for example:
1.
How to search for details in an image?
How do facial recognition systems work?
Are there open source neural networks for finding faces in photographs?
A system for detecting an object in a photo?
2
. Is it possible to determine the distance to the object from the picture?
Does the distance to the object affect the blur?
Are there any programs for calculating the distance to an object using a webcam or a smartphone / tablet?
https://bootsector.livejournal.com/43436.html
And so on.
"how to implement it."- Well, after reading the links above, there will also be answers.
What are self-learning algorithms other than "perceptron".
Well, for example:
https://habr.com/ru/post/277145/
Only pampering is all. For the love of art, you can look. Real use - is unlikely to succeed.
And in general, your question began with the words: "I have an idea." . No idea anywhere?