Made in 2 days. I tried to create a facial recognition system, since the front page ones are notoriously inaccurate. I had to solve some critical flaws first: Since the camera feature is a bit inaccurate, this system excels by taking multiple snapshots of the face, observing the trends, and averaging the patterns into a stable biometric profile. Additionally, I had to address face posture prior to picture, where I came to realise that calculating face yaw and pitch offsets was necessary, Recognition calculates a weighted percentage of similarity based on stabilised capture data, which passes a threshold, determining match status. This works reliably with stable lighting and camera distance. Credit is required for use. > Update Log - (v1) Initial system, incorporated all features listed above. - (v2) Added face profiling for more facial data, multiple fixes, improved UI, added advanced setting modification, cosmetic improvements, and more.
