Emre
  • Member for 9 years, 11 months
  • Last seen more than a month ago
How to distinguish voice from snoring?
11 votes

Background According to papers below, snoring is characterized by a peak at about 130Hz, and is wholly concentrated below 12kHz: Non-invasive Sensors based Human State in Nightlong Sleep Analysis ...

View answer
How to Remove Gaussian Noise from an Image without Destroying the Edges?
10 votes

You might need to consider more advanced techniques. Here are two recent papers on edge-preserving denoising: Edge-Preserving Image Denoising via Optimal Color Space Projection [in color] This paper ...

View answer
Who is Barbara (test image)
8 votes

The appeal of this image is obviously in the numerous lines, which test the aliasing properties of resizing, denoising, and super-resolution algorithms. It seems Allen Gersho is the source, according ...

View answer
Digital image moments in plain English
8 votes

An image moment is simply a number that characterizes the image, construed as the realization of a spatial random variable. If you've taken any probability class, you should remember the concepts of ...

View answer
Identify the correct spot to place a label
8 votes

I submit that the ideal place to place the label should meet two objectives: proximity to the center, say $d$. legibility, say $l$. Ergo, we can determine the ideal point by minimizing a holistic ...

View answer
What does frequency domain denote in case of images?
8 votes

In images increasing frequency is associated with more abrupt transitions in brightness or color. Furthermore, noise is usually embedded in the high end of the spectrum, so low-pass filtering can be ...

View answer
Conceptually, how does a Fourier transform differ from an autocorrelation?
7 votes

For starters, autocorrelation is a function of the relative time only for WSS processes, otherwise it depends on the absolute times: $\mathrm R_X(t_1,t_2) \equiv \mathbb E[X(t_1)^* X(t_2)]$ Secondly, ...

View answer
How best to digitize a noisy phone recording for filtering?
Accepted answer
7 votes

Unless you have tight space constraints, I would stick to lossless formats like good old WAV. Phones do not pass much high-frequency content so you should be fine with a 8KHz Mono WAVE file. You will ...

View answer
Why is telephone audio sampled at 8 kHz?
6 votes

It was thought to provide a good trade-off between quality and bandwidth. Actually a single voice signal occupies 8 kHz, not 8 kbps, of bandwidth. Each sample is quantized into 8 bits, yielding a rate ...

View answer
Image comparison using MATLAB
6 votes

You can borrow from techniques used by the image stitching community, where they have to deal with registering translated and rotated images. Your mistake was that you calculated the correlation ...

View answer
Otsu binarization with value preservation
6 votes

You can easily do it with Mathematica: img = ColorNegate@ColorConvert[Import["../Desktop/sample.jpg"], "Grayscale"] ColorNegate@ImageMultiply[Binarize[img], img] The negation and multiplication ...

View answer
What's the unit of spectral intensity in calculation of RGB tristimulus?
Accepted answer
6 votes

You are probably working not with CIE RGB, but CIE XYZ, so these are properly denoted the XYZ tristimulus values, not RGB. Their absolute values do not matter if you want to determine the ...

View answer
A non-lossy aka lossless codec without patent-restrictions for audio?
Accepted answer
6 votes

FLAC is the de facto standard, particularly, notice from the latter link: The specification of the stream format can be implemented by anyone without prior permission (Xiph.org reserves the right ...

View answer
I trigger recording when power of current sample is X dB greater than last Y samples, what is the terminology here?
Accepted answer
6 votes

"Running average" is fine. You can also call it the moving average. I suppose you could call the variable associated with the 10dB number the signal detection threshold.

View answer
How do you calculate spectral flatness from an FFT?
5 votes

Definitions vary, don't they? The first thing that has to be settled is whether we agree that the power spectral density is equivalent to the power spectrum, or else define what we mean by both. ...

View answer
What does it mean to combine signals coherently?
5 votes

It means that they are in phase---think of the signals as waves. Adding noncoherent signals results in cancellation, or fading.

View answer
Intuitive explanation of tracking with Kalman filters
5 votes

The Kalman filter recursively provides the optimal linear estimate of a signal perturbed by AWGN. In your case, the state (what you want to estimate) will be given by the target location. The ...

View answer
Implementing Gaussian random variable by using a uniform random variable
5 votes

how can I change variance of this random variable? By multiplication, of course. The variance of $c X$, where $c$ is the multiplicative constant and $X$ is your random variable, is $c^2$ the variance ...

View answer
How to mimic/copy/fake someone's voice?
4 votes

You can use something like MorphVox. Here is a demonstration. The process is called voice morphing or conversion. If you are interested in the technical aspects, a recent article you can study is ...

View answer
A signal containing no pure sinusoidal components
Accepted answer
4 votes

It contains no spikes in its frequency spectrum. (What is the Fourier transform of a sinusoid?)

View answer
Frequency spectrum of a sinc function
3 votes

Look at the forms of the functions, and remember the duality property. I invite you the read the Wikipedia article: http://en.wikipedia.org/wiki/Fourier_transform

View answer
Functional analysis for signal processing engineers
3 votes

It depends on whether you intend to enter academia or industry. Academics write papers and can use all the maths they can get. In industry not so much, unless you intend to enter a research lab, which ...

View answer
Getting stuck into an open source image processing/computer vision project
3 votes

I believe it would if you intend to practice your knowledge, rather than go into academia. I suggest you look into the code base of opencv, simplecv, and scikit-image, as suggested. Once you become ...

View answer
When should I use a frequency spectrum rather than a spectrogram representation?
3 votes

You use a spectrogram (time-frequency analysis) if you need to know when a frequency component is active, and a frequency spectrum when you don't. The drawback of using a spectrogram is that it has ...

View answer
Best way to evaluate "quality" of autocorrelation?
3 votes

Preliminaries This demonstration is easier with the MATLAB package audioread, which enables reading/writing MP3 files. Alternatively, you can convert the MP3 file in the example to WAV manually. ...

View answer
What are some face detection (not recognition) algorithms suitable for limited (embedded) processors?
Accepted answer
3 votes

We know that the problem is solved because our cell phones and digital cameras can do it. You might be able to find out what algorithm they use by trawling the patent databases. Computer vision ...

View answer
how to remove color of light source?
3 votes

The operation you need to perform is called white balance or color balance. Several algorithms are described in the article. An important proviso is that the light source be approximately white, not ...

View answer
Extract repeating structure from an image
2 votes

You want to detect two things: lines/curves and circles. There are well-known algorithms to do both, so let's find them separately. I will present a solution in Mathematica. You can translate to your ...

View answer
how can we tell that OTSU binarization is feasible or not.
Accepted answer
2 votes

Assuming you have MATLAB, you can just look at the source code, where you will see that the EM is defined as em = maxval/(sum(p.*((1:num_bins).^2)') - mu_t^2); where I = im2uint8(I(:)); # the ...

View answer
Extract graph data from map image
2 votes

Start by getting rid of the background (Mathematica example): map = Import["http://i.stack.imgur.com/HYtbo.jpg"] fg = ImageAdjust[map, {1.0, 1.0}] Then use a graph extraction tool like NEFI. ...

View answer