How do we choose d? Is there a standard value for d?
ellahofm
If you reduce the original image by a factor of two till you have only one pixel in either height/width (or both in the case of a square image), then d = log_2(n) where n is the min(height,width) of an image. I'd imagine d would also potentially be adjusted to fit a storage constraint of the resulting mipmap.
ehsan
this reminds me of wavelet compression (like what is used in JPEG images)
How do we choose d? Is there a standard value for d?
If you reduce the original image by a factor of two till you have only one pixel in either height/width (or both in the case of a square image), then d = log_2(n) where n is the min(height,width) of an image. I'd imagine d would also potentially be adjusted to fit a storage constraint of the resulting mipmap.
this reminds me of wavelet compression (like what is used in JPEG images)