Just for fun, here's how some of OpenCV's feature detectors handle the first image: AKAZE: https://imgur.com/EvcStcH https://imgur.com/EvcStcH ORB: https://im

Just for fun, here's how some of OpenCV's feature detectors handle the first image: AKAZE: https://imgur.com/EvcStcH https://imgur.com/EvcStcH ORB: https://imgur.com/YHSEDD7 https://imgur.com/YHSEDD7 I certainly wouldn't make any claims about what this says about the human visual system, but at least for these two algorithms they do seem to have troubles finding interesting key points to track on the buildings.