4
The answer can be given without being at all specific to filters. In fact, one has to answer that in terms of engineering in general:
Whenever you have alternative solutions, the only measure of quality that really matters, is fulfillment of the requirements.
So, you'll have to know what you'll use that filter for, and evaluate how well the alternatives ...
2
The method to improve resolution by few frames is called Multi Image Super Resolution (As opposed to Single Image Super Resolution).
For compression artifacts you can look for JPEG Deblocking algorithms.
2
Interesting code! This was tough to follow from the pictures, but I did run the OP's code from the linked site, I can't play the sound loud enough to really test but appears to not be very sensitive for lower volume signals. There is no indication of the frequency axis on the graphic but I assume the left side of the screen is DC and the right side of the ...
1
As video compression implicitly tries to track scene motion and express the residual approximately in a compact (and visually pleasing manner), I think that applying general multi frame super reolution is unlikely to give good results unless your video stream is encoded using only i-frames (eg motion jpeg, where temporal information is not exploited by the ...
1
Beside Video Super-Resolution, as answered by @Royi,Other words for this task (fill in missing regions of a given video sequence with contents that are both spatially and temporally coherent) are video inpainting, video completion, or even video restoration.
There are many flavors: motion-based, object based, using deep learning, etc. A couple of links
...
1
TL,DR: the low-pass component (approximation coefficients, a) has a size bigger than expected ($2^3$ times). So I guess that the 8 avatars of the approximation subbands are gathered into one.
First, I did not take enough time to check the codes, so this may be a partial answer. While performing a dualtree3 decomposition:
zr = rand(64,64,64);
[a,d] = ...
1
Prof. Nick Kingsbury kindly provide an answer to my question!.
In 1-D, the lowpass basis functions (scaling functions) from the two
trees (a) and (b) of the dual tree WT tend to look very similar to
each other, apart from a shift of half the output sample period
between them. Hence it usually makes most sense to regard the two sets
of lowpass samples as ...
1
Yes, they can be different, depending what your maximum input value is.
1
A "quantization process" is a mathematical abstraction.
Things like analog to digital converters, conversions from wider databusses to narrower databusses, etc., are actual physical processes.
A purely mathematical process that quantizes in steps of 2 (i.e., $\left \lbrace \cdots, -3, -1, 1, \cdots \right \rbrace$ would just keep quantizing, ...
1
In a typical quantization like that seen in an ADC the output would saturate any value > 3 to 3 (also known as clipping). Note that if you're considering a software quantization, it's also possible that some quantizers (although IMO poorly designed ones) could cause out of range values to wrap around and switch signs.
1
to obtain the desired ideal filter.
That's like asking: how do design the ideal car? There is no such a thing: The reason why there are so many different car models, is because people have very different needs and requirements. Each car model represents a different trade off between size, cost, efficiency, loadability, towing ability, aesthetics, ...
Only top voted, non community-wiki answers of a minimum length are eligible
