|
||||||||||||||||||||||
|
||||||||||||||||||||||
In the previous section we do not use any filter to reduce high frequency component of the image. This may result in aliasing effect in the subsampled images which leads to a problem in progressive transmission.
To illustrate this effect, consider the following extreme image. |
||||||||||||||||||||||
|
||||||||||||||||||||||
This is an image of a sine wave with increasing frequency from top left to bottom right. Next consider the images in the Gaussian pyramid in the previous two sections. | ||||||||||||||||||||||
g1 | g2 | g3 | ||||||||||||||||||||
|
||||||||||||||||||||||
g1 | g2 | g3 | ||||||||||||||||||||
|
||||||||||||||||||||||
Notice that linear low-pass filter gives less aliasing especially at higher level in the pyramid. Now we see that each filter has its own advantage and disadvantage. In the next section, we shall introduce a method to combine these filters. | ||||||||||||||||||||||
|