I want to build a program in C++ (I use Visual Studio 2019) using OpenCV libraries. Here is the program definition.
- I have a .jpeg (call it “jpeg A”) with an object (call it “Object A”) in the very center of jpeg A.
- 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.
- I have another .jpeg (call it “jpeg B”) that may or may not contain an object similar (in size & shape) to object A.
- Build a program what will output the probability that a similar version of Object A is somewhere in jpeg B.
- 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?