After reading text about this said topic, i found out that it considers 16 of the original neighboring pixels. What i want to know is how does it compute the color value of the new pixel. If the color of pixels can be determined by 200,100,255, how could you compute the value of the new one?
+8
A:
I think it's pretty well explained in Wikipedia. You need the intensity values of 4*4=16 pixels, from which you can calculate the interpolated value at any point within that 4*4 grid.
If you mean how to do this for RGB triplets, you just do the process separately for each component.
Joonas Pulakka
2009-12-30 08:42:46
what i need is pixel level computation. can you give me a concrete example
aidriiyan
2010-01-03 13:50:34
+1
A:
You should spend some time learning how to do a bilinear interpolation, as a tensor product linear interpolation. Once you understand it and how/why it works, then the bicubic case is easier to follow.
woodchips
2010-01-01 03:39:58
correct me if im wrong.(bilinear) Example(1 intensity only): p00=255, p01=0, p10=100, p11=255new pixel= [((255+0)/2) + ((100+255)/2)]/2= 153.... am i right?...if yes, how could i implement it in bicubic? if no, give me a good example.
aidriiyan
2010-01-03 13:48:56