# Can aliasing happen in Laplacian Pyramids?

I want to use the Laplacian pyramid framework for representing images. The The Laplacian Pyramid as a Compact Image Code's language seems to provide freedom in the choice of blurring/interpolation kernel, subject to certain constraints. However, as we are performing subsampling when constructing pyramids, would aliasing be a problem with an injudicious choice?

I read in the Toronto lecture notes on image pyramids(Slide 8) that a laplacian pyramid can reconstruct an image exactly, agnostic of the kernel. Can i take this to be a guarantee against aliasing, or can aliasing be "hidden" by the reconstruction process?

• Yes it can makes aliasing. If you want to remove the aliasing then you sould find the frequency what itself. That is, you can divide it by Nyquest theory. – gmotree Jun 8 '15 at 9:01
• Your question has beeen answered. Do not hesitate to vote for the useful ones and accept the most suitable – Laurent Duval Feb 9 '17 at 17:25

Let's look at the first iteration of the encoding process. The input image is lowpass filtered and then downsampled (decimated) to obtain the lower resolution image that will function as the input to the next iteration. Because the lowpass filter is not ideal, some aliasing will take place during downsampling. If the prediction error (to be stored in the encoding) is calculated not from the lowpass filtered signal but from an upsampled version of the lower resolution image, then it will also contain the aforementioned aliasing error. Decoding (reconstruction) will then also correct for the aliasing error.

• Yes these were the two processes i was considering: aliasing due non-ideal low pass filtering, and the reconstruction hiding this aliasing. But what I'd appreciate is an explanation of how the aliasing is hidden/ corrected during reconstruction. AFAIK aliasing is a lossy step where the signal self overlaps and loses information, so how is it possible to correct for it. – user3246971 Jun 8 '15 at 4:55
• Let's assume for the sake of conversation that instead of lowpass filtering and downsampling we do something ultimately destructive: zeroing all pixels. Still there is no loss as the encoded "prediction error" includes error due to zeroing, for each and every pixel. – Olli Niemitalo Jun 8 '15 at 7:20
• As the question was lacking activity, and just bumped this morning, let us cast some votes – Laurent Duval Dec 29 '16 at 9:26

Yes, most of the known multiscale or multirate decompositions, as long as they combine at one stage a non-ideal filter and a subsampling operator, induce some kind of aliasing at the analysis stage. And the Laplacian pyramid does so. But proper designs allow perfect reconstruction, so the aliasing can be reversed or cancelled, and kept reasonable in the case of processing between analysis and synthesis.

The simplest form of alias cancellation arises with Haar sum and difference filters. Consider $h_l = [1,1]$ and $h_h = [1,-1]$, the sum and difference filters, followed by a 2-fold downsampling. Both filters have very poor frequency responses, and cause aliasing with downsampling.

Take signal $X=[a,b]$. With the above scheme, you get two outputs $y_l = a+b$ and $y_h = b-a$, from which you can recover $X$ with $(y_l-y_h)/2 = a$ and $(y_l+y_h)/2 = b$. The latter operations are filtering by an averaged difference filter, and an averaging filter, e.g. the same as above, with a different scaling factor.

That is the magic of perfect reconstruction filter-banks: from well-chosen imperfect analysis filter-banks (with some decimation), to be able to design a synthesis filter-bank, itself imperfect, that cancels aliasing and distorsion of the whole analysis-synthesis system. It may happen as long as you have two or more channels, and with a little oversampling in the system, you can even find optimized inverse filter-banks.