I am testing the efficacy of the 3 facerecognizer algorithms openCV proffers:
Eigenface Fisherface LocalBinaryHistogramPattern
And believe the latter is the best so far. I cannot find an adequate explanation for how the initialization parameters affect this approach. They are
radius, neighbors, grid_x, and grid_y
I can intuit some of it but the task of sussing them out by changing them little by little and observing is eating up time I don't have. If anyone knows of a detailed explanation of them, I would be most grateful.
Thanks and I hope this hasn't been answered elsewhere - I tried looking but with no success...
Sean