It has been 20 years since I first encountered homography as a graduate student in 2005. Throughout my Ph.D. studies (2007–2012) and subsequent university teaching career (2013–present), homography has remained my primary research focus, continuing to hold mysteries for me. Although it took over a decade after my most recent previous work (published in IVC in 2017, submitted in 2015) to develop the SKS and ACA homography decomposition methods, I am still uncertain whether our approach represents the optimal solution for 4-point homography computation.

To encourage further advancements in the efficiency of 4-point homography computation, I am personally offering two tiers of cash rewards:

Minor Improvement: A prize of USD 100 will be awarded for the first method, or any totally different method, that reduces the number of floating-point operations (FLOPs) for general 4-point homography computation by at least 2 compared to our ACA method (i.e., achieving ≤83 FLOPs).

Major Improvement: A prize of USD 1000 will be awarded for either:

1. Reducing the FLOPs for general 4-point homography computation by at least 10% compared to our method (i.e., achieving ≤77 FLOPs).

2. Reducing the FLOPs for square pattern's homography computation by at least 10% (i.e., achieving ≤26 FLOPs).

If you have any questions or suggestions about rewards, or ideas to extend our SKS/ACA method to other vision, graphics, or mathematical problems, please don't hesitate to contact me via email. Our ongoing research might align with your interests, and I welcome discussions and collaborations.