This content originally appeared on DEV Community and was authored by 이상우
Hello, this is sangwoo from Korea
and I have a problem. I am currently making the AI algorithm that have to determine if the building or bridge are actually bent or not, when I took the photos of them.
the image above, let’s say right side is original photo data and left side is new photo data, I have to judge the curved part of the image on the left within the matched area. However, in fact, the building is not bent, and only the angle and location of the shooting have been changed.
but the above is the result of the my AI code. it just recognised it as it’s bent. I don’t have a concept that I can define that ‘No, that’s just moving. It’s not bent.’
I am struggling with issue more than a months, please someone helps me. thank you.
This content originally appeared on DEV Community and was authored by 이상우