bitmap interpolation

Question

位图的插值：我有16 * 16的位图，我想将位图的大小增加到160 * 160，这是最适合的插值类型。

Answer 1

• Bicubic interpolation (cubic spline) blurs area and because of that destroys edges.
• Nearest neighbour interpolation preserves edges, but introduces pixelation.

So best interpolation of bitmap would be hybrid algorithm of those two above - such as 2xSal / Eagle and such .

EDIT: Bicubic interpolation JAVA example code .

Good luck.

Answer 2

Um, this is a horrendously bad idea no matter which way you look at it - you're basically increasing the size of the bitmap by an order of magnitude, but you're not adding any new data - regardless of whether you use a polynomial, linear, or simple copy mechanism the result is going to show either extreme pixelation, extreme blurring, or some mixture of the two.

In general, these algorithms work much better for reducing the size of bitmap images and maintaining overall image integrity, blowing an image up by an order of magnitude is going to be ugly, no matter what you do - the bottom line is that the original information is now dwarfed by the information the interpolater 'made up'