Why do you use inverse H?

[giovannicaprio]

Dear @alexhagiopol , I am confused, could you explain to me why you train using your inverse H instead of the H itself?

thank you Giovanni

[huangchaoxing]

Hi @giovannicaprio This is because we still crop from point P in the new image. We have warp P to a new destination point in the image using H and we want to know the grayscale value of this destination point without knowing the actual pixel coordinate of the destination. Therefore, we use H inverse to warp the destination point back to P and we can assign the new grayscale value to P, since we still crop from P. Hope it helps.

Chaoxing

[giovannicaprio]

Hi @huangchaoxing yes it is very useful. Thank you

Giovanni