Is this feasible with OpenCV libraries?

I want to build a program in C++ (I use Visual Studio 2019) using OpenCV libraries. Here is the program definition.

  1. I have a .jpeg (call it “jpeg A”) with an object (call it “Object A”) in the very center of jpeg A.
  2. Object A is defined by a number of pixels but could be defined within a certain size box if necessary. Object A contains at least 20% of all the pixels in jpeg A.
  3. I have another .jpeg (call it “jpeg B”) that may or may not contain an object similar (in size & shape) to object A.
    The Goal:
  4. Build a program what will output the probability that a similar version of Object A is somewhere in jpeg B.
  5. Good accuracy is important but could be as low as ±20%.
    Example of the output:
    0% or 43% or 82% or 100%.
    Is this feasible with OpenCV libraries?


that depends entirely on what you consider an “object”. your description couldn’t me more abstract. you’ll have to show concrete pictures of your precise situation.

Your reply fully answers my question.