Signal Processing - Science topic

It will be a great moment to contribute on how neural network can be more efficient in the domain of Information technology Sijia Ma Sijia Ma

Look here:

EMG is low-frequency signal bursts with high-frequency noise-like filling. Most often, the amplitude of the bursts and the repetition rate are analyzed. And the high-frequency noise-like filling is probably not very informative.

But in EEG, they conduct a spectral analysis of a noise-like signal with a slowly changing envelope in different frequency filters, which are associated with specific rhythms.

In ECG, little attention is paid to frequency spectra, but the period and shape of the signal are analyzed.

Unfortunately, I can't read the source

in the original language.

A rate-distortion optimization algorithm based on visual perception

A sliding-clustering-based method for multi-sensor asynchronous information fusion

Non-contact two-dimensional haptic rendering system based on electromagnetic force control

Point cloud simplification and reconstruction parameters’ automatic adjustment method of structured light detection

A position and attitude calibration method for the linear laser sensor in gear 3D measurement

Hi Isaac Agyapong ,please see the official website of the conference, which is in English: https://ais.cn/u/BBfQBb

Hello Prof Valerie Renaudin, Prof Joaqin Torress, Dr Pampa Sadhukhan

This topic sounds interesting. It has a relationship to a current/ongoing research project, building a submersible robot to explore deep lakes and open ocean marine environments. The goal is to build a device that will be self navagating containing obstacle detection and inherent collision avoidance as well as the obvious wide ranging data collection capabilities. I aim to make the robot fully autonomous with no tethering to a surface vehicle required. The functional model would be a kind of "Throw it over the side and let it explore", mode of operation.

Would it be possible to keep in touch with you folk to see the end results of these investigations as there may be emerging interesting adaptations that are applicable to my project.

Regards

Murray Foote

definitely yes

The most modern approaches in images' denoising are based on machine learning methodologies.

You can try publishers with agreements and discounts with your institution to cover the APC

Dear Bouziane Ghoual ,Thank you for your attention to EEI 2024. We apologize that considering the on-site experience, this conference only accepts offline presentation in China.

If you are interested in this conference, you could consider submitting your papers and attending the conference offline. Or you could also consider joining as a listener without submmison online.(Listener could participate online）

Useful thing

Take look on

Jacob Sundstrom Just curious if you still have the zip file.

It seems that the acquisition is done in zigzag manner which created the strips of variation in Magnetic field. To remove this heading error use software MAGMAP.

File-Destripe data wizard - choose input .dat file and give a name for output file-click next- select channels which need to be filtered--click Next twice-give line smoothing value 1-2 depends upon the data- press Next and Run Destriping.

Done.

You can check the attached manual. Article 7.5, page no. 116.

Hey there S M Mohiuddin Khan Shiam! Convolution in signal processing is like the secret sauce that helps us unpack the mysteries of linear time-invariant (LTI) systems. Picture it as the detective work for signals – it reveals hidden patterns and relationships.

So, what's the deal with convolution and LTI systems? Well, buckle up. When we apply convolution to signals, it's like taking a signal and sliding it over another while computing the integral of their product at each point. This process highlights how the input signal influences the output, and it's gold for understanding LTI systems.

Why is it crucial? Imagine you're dealing with a system that doesn't change over time, like a stable filter or circuit. Convolution helps us predict the system's response to any input, making it a cornerstone in system analysis. It's like having a crystal ball for signal behavior.

Applications? Everywhere. From image processing to audio filtering, convolution's fingerprints are all over. It's a powerhouse in understanding how systems react to different inputs, giving engineers the upper hand in designing and optimizing systems.

In a nutshell, convolution in signal processing is the Sherlock Holmes of understanding linear time-invariant systems. It unveils the hidden connections and intricacies, making it an indispensable tool in the engineer's arsenal.

thank you for the information

I have observed after pandemic, there are open access journals. So, editors suggest to submit in the open access journals. As a result, we are facing this issue. It can be one of the reason. Solution to this problem from my point of view:

1) Target IEEE Journals (there are fare chances of acceptance here)

2) Focus on experimental research work

3) Focus on patents based on original research work

Murtadha Shukur thank you for this clear answer

Dear if you are looking for Radio Bearer AT Commands for UMTS then here is the link that may help you please. https://www.sparkfun.com/datasheets/Cellular%20Modules/AT_Commands_Reference_Guide_r0.pdf

Here's a follow up Camille,

Weight Matrix w in TRF Analysis:

The weight matrix w is a fundamental output of Temporal Response Function (TRF) analysis, providing insights into how different aspects of the stimulus relate to neural activity.

Mathematical Representation:

- Each row of w corresponds to a specific time window in the stimulus, denoted as t=1, t=2, t=3, and so on.

- Each column of w corresponds to a neural activity channel, represented as Channel 1, Channel 2, and so forth.

- The values in the weight matrix w are calculated using the formula:

w = (STS)^-1STr

Example:

Suppose we have a simplified weight matrix w, where rows represent different time windows and columns represent neural channels:

| w1, Channel 1 w1, Channel 2 ... w1, Channel N |

| w2, Channel 1 w2, Channel 2 ... w2, Channel N |

| w3, Channel 1 w3, Channel 2 ... w3, Channel N |

In this matrix:

- w1, Channel 1 represents the weight or correlation between the first time window (t=1) of the stimulus and neural Channel 1.

- w2, Channel 2 represents the weight or correlation between the second time window (t=2) of the stimulus and neural Channel 2.

- Each value w captures how strongly a specific time window influences the activity in a particular neural channel.

Interpretation:

- Larger positive values of w indicate that a particular time window has a strong positive influence on the neural activity in a given channel.

- Smaller positive values indicate a positive but weaker influence.

- Negative values suggest a negative correlation, meaning that the time window has an inhibitory effect on neural activity in that channel.

Practical Use:

By examining the weight matrix w, researchers can pinpoint which temporal aspects of the stimulus are most relevant for explaining neural responses. This information is crucial for understanding how auditory stimuli are processed in the brain and aids in the decoding of auditory speech perception.

Your '-160' must be below -180 as you pass simple inversion, I would say -240.. Also, your '+20' is more like 120 degress or so (90 degree shift = top coninceds with 0 crossing).

120+240 =360.

One should note it is 360 deg (full period ) between sinusoiidal peaks, and you may choose to represent phase as 0..360 or +-180 degress

DTFS (Discrete Time Fourier Series) and DFT (Discrete Fourier Transform) are mathematical tools used to analyze discrete-time signals in the frequency domain. Although they are both related to the Fourier transform, they differ in their domain and representation.

DTFS is used to represent a periodic discrete-time signal in the frequency domain. It is defined as the Fourier series representation of a periodic sequence of samples. DTFS is useful when analyzing signals that are periodic in nature, such as audio signals or digital signals with a fixed frame rate. DTFS coefficients represent the frequency components of a periodic signal and are discrete in nature.

DFT, on the other hand, is used to represent a finite-length discrete-time signal in the frequency domain. It is defined as the Fourier transform of a finite-length sequence of samples. DFT is useful when analyzing signals that are non-periodic in nature, such as speech signals or biomedical signals. DFT coefficients represent the frequency components of a finite-length signal and are also discrete in nature.

The main difference between DTFS and DFT is their input and output. DTFS requires a periodic signal as input and produces a set of discrete frequency components as output. DFT requires a finite-length signal as input and produces a set of discrete frequency components as output. Another difference is that DTFS coefficients are complex, while DFT coefficients are also complex but are usually represented as real and imaginary parts.

In summary, DTFS is used to represent periodic signals in the frequency domain using discrete frequency components, while DFT is used to represent finite-length signals in the frequency domain using discrete frequency components.

In signal processing, the sensitivity to noise detection can vary depending on the specific application and the characteristics of the noise and signal being processed. However, in general, a second-order operator tends to be more sensitive to noise detection compared to a first-order operator.

This steeper roll-off characteristic allows the second-order operator to attenuate higher-frequency noise components more effectively compared to a first-order operator. Therefore, a second-order operator can be more sensitive to noise detection because it can provide better noise reduction, especially for higher-frequency noise.

It's important to note that the specific characteristics and requirements of the signal and noise in a given application can influence the choice of the operator.

Wavelet Transform and Hilbert Huang Transform (or Empirical Mode Decomposition) are suitable for non-linear and non-stationary signal processing and analysis. 1st one depends on choice of basis or mother wavelet and 2nd one is a data-adaptive method.

Anyone can evaluate it.English is not my native language, so I hope there is no ambiguity.

Only a limited number of samles are displayed on the screen, possibly less than the length of the record.

The duration of a record is determined by its length and the sampling rate.

The duration of the phenomena on the screen is determined by the width of the screen, the sampling rate and the scale of the screen.

For optimal processing, the length of the record must be matched to the physical characteristics of the recording medium (for example, the block size on the disk).For optimal processing, the length of the record must be matched to the physical characteristics of the recording medium (for example, the block size on the disk).

Thanks for answering Aparna Sathya Murthy

Please read my paper

An Adaptive Filter to Pick up a Wiener Filter from the Error using MSE with and Without Noise

This is a system that is able to learn.

The paper is a singles and systems.

The topic is AI.

I think the two fields support each other.

Thank you

Ziad

Join us at http://icit.zuj.edu.jo/icit2023/Index.html

From your description, I suppose you know the random signal value (denoted by r(t)) within the time interval from 0 to T=0.01 seconds, but don't know (or, do not care) the signal value outside this interval. In order to approximate this signal as a superposition of harmonic signals, you may try the following.

1) Define a signal s(t) which is equal to r(t) within the interval from 0 to T, but is zero outside this interval.

2) Define a signal x(t) which is a sum of s(t-kT), where k is integers from minus infinity to infinity.

In this way, x(t) is an infinitely long periodic signal with its value same as the random signal r(t) within the interval from 0 to T, and, it can be expressed as a superposition of harmonic signals since x(t) is periodic.

Hope it helps. Good luck.

The Fourier transform of a real signal is symmetric; hence the mirror image. To ommit this redundancy, we use the analytic version of the signal, which is complex, and results in positive frequency spectra

There are many estimation techniques that can be used to estimate system poles. Here are a few popular ones:

To apply an algorithm to estimate system poles, you can start with a simple transfer function and apply the algorithm to estimate the poles. You can then compare the estimated poles with the known poles of the transfer function to evaluate the accuracy of the algorithm. It may also be useful to test the algorithm on more complex systems to see how well it performs.

[1] Bazzi, Ahmad, Dirk TM Slock, and Lisa Meilhac. "Efficient maximum likelihood joint estimation of angles and times of arrival of multiple paths." 2015 IEEE Globecom Workshops (GC Wkshps). IEEE, 2015.

[2] Bazzi, Ahmad, Dirk TM Slock, and Lisa Meilhac. "On a mutual coupling agnostic maximum likelihood angle of arrival estimator by alternating projection." 2016 IEEE Global Conference on Signal and Information Processing (GlobalSIP). IEEE, 2016.

[3] Bazzi, Ahmad, Dirk TM Slock, and Lisa Meilhac. "On Maximum Likelihood Angle of Arrival Estimation Using Orthogonal Projections." 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2018.

Yes, here are some potential research topics on smart cities that you could explore:

Certainly, here are some articles and book chapters you can refer to:

These resources should provide you with a good starting point to understand Total Variation in classical signal processing and its relation to Graph Signal Processing and Fourier Basis.

To filter an input signal through a lognormal shadowing model or kappa mu shadowing model and generate a probability density function (PDF) in MATLAB, you can follow these steps:

Here's an example of how you might use this process for the lognormal shadowing model:

% Generate the input signal

fs = 1000; % sampling frequency

t = 0:1/fs:1-1/fs; % time vector

x = sin(2*pi*100*t) + sin(2*pi*200*t) + sin(2*pi*300*t);

% Define the parameters of the shadowing model

mu = 0; % mean of the underlying normal distribution

sigma = 1; % standard deviation of the underlying normal distribution

% Generate the shadowing factor sequence

shadowing = exp(mu + sigma*randn(size(x)));

% Filter the input signal

y = x .* shadowing;

% Generate the PDF of the filtered signal

nbins = 100;

histogram(y, nbins, 'Normalization', 'pdf');

xlabel('Signal Amplitude');

ylabel('Probability Density');

title('PDF of Filtered Signal');

==================

In this example, the input signal is a sum of three sine waves, and the lognormal shadowing model is used with a mean of 0 and a standard deviation of 1. The "shadowing" sequence is generated by taking the exponential of a sequence of random variables from the underlying normal distribution. The input signal is then filtered by element-wise multiplication with the "shadowing" sequence, and the PDF of the filtered signal is generated using the built-in "histogram" function.

To maximize the log-likelihood estimate of a signal using maximum likelihood estimation in MATLAB, you can use the built-in optimization functions. Here's a general process you can follow:

So it depends on the model you have at hand. Here's some papers applying MLE for different type of problems[1-3]:

[1] Bazzi, Ahmad, Dirk TM Slock, and Lisa Meilhac. "Efficient maximum likelihood joint estimation of angles and times of arrival of multiple paths." 2015 IEEE Globecom Workshops (GC Wkshps). IEEE, 2015.

[2] Bazzi, Ahmad, Dirk TM Slock, and Lisa Meilhac. "On a mutual coupling agnostic maximum likelihood angle of arrival estimator by alternating projection." 2016 IEEE Global Conference on Signal and Information Processing (GlobalSIP). IEEE, 2016.

[3] Bazzi, Ahmad, Dirk TM Slock, and Lisa Meilhac. "On Maximum Likelihood Angle of Arrival Estimation Using Orthogonal Projections." 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2018.

Selecting the right signal processing method for damage detection in structures can be a challenging task, as different methods may be better suited to different types of damage and structural systems. Researchers typically consider several factors when selecting a signal processing method, including the characteristics of the signal being analyzed, the type of damage being detected, the sensing and measurement system used, and the computational resources available. Here are some of the factors that researchers typically consider when selecting a signal processing method for damage detection in structures:

Overall, selecting the right signal processing method for damage detection in structures requires careful consideration of the characteristics of the signal being analyzed, the type of damage being detected, the sensing and measurement system used, and the computational resources available. By taking these factors into account, researchers can select a method that is well-suited to their specific application and achieve accurate and reliable damage detection results.

Digital filtering involves the processing of digital signals to achieve a desired signal response. There are several technical challenges that need to be considered when designing and implementing digital filters. Here are some of the common technical challenges in digital filtering:

Overall, digital filtering is a challenging task that requires a thorough understanding of the filter characteristics, the signal processing requirements, and the computational resources available. By carefully considering these factors, it is possible to design and implement digital filters that meet the specific needs of the application.

Hermitian signal processing is a type of signal processing that is used to analyze signals with Hermitian symmetry. Hermitian symmetry is a property of complex-valued signals where the real and imaginary parts of the signal are symmetric around the center of the signal. Hermitian signal processing takes advantage of this symmetry to reduce the computational complexity of certain signal processing operations, such as Fourier transforms.

To perform Hermitian signal processing in MATLAB, the following steps can be taken:

Note that the result of the Hermitian symmetry operation is a real-valued signal, since the imaginary components of the FFT result are negated and combined with the real components to form the final result. The real-valued signal can then be used for further signal processing operations, such as filtering or spectral analysis.

Daubechies wavelet filters are a family of orthogonal wavelet filters that are commonly used in signal processing applications. The primary function of Daubechies filters in wavelet transforms is to analyze and decompose signals into their constituent wavelet components, which can reveal valuable information about the signal's time and frequency characteristics. The wavelet transform is a mathematical tool that decomposes a signal into different frequency bands, with each frequency band represented by a set of wavelet coefficients. Daubechies filters are used to analyze and decompose the signal into different frequency bands, by applying a series of high-pass and low-pass filters to the signal.

The Daubechies wavelet filters have several desirable properties that make them particularly useful in signal processing applications. For example, they are orthogonal, which means that they can accurately represent both the time and frequency characteristics of a signal. They also have a compact support, which means that they only require a finite number of filter coefficients, making them computationally efficient. Regarding signal decomposition, Daubechies filters can also be used for signal denoising, feature extraction, and compression. In signal denoising, wavelet coefficients with low magnitude are truncated or set to zero, effectively removing noise from the signal. In feature extraction, wavelet coefficients are used to extract relevant information about the signal, such as the location of edges or sharp transitions. In compression, wavelet coefficients are used to represent the signal in a more efficient and compact form, which can reduce storage and transmission requirements. Moreover, Daubechies wavelet filters are an essential tool in signal processing applications, allowing for accurate and efficient analysis, denoising, feature extraction, and compression of signals.

Geophysical signal processing is an important field that deals with the analysis, interpretation, and modeling of various geophysical data, such as seismic, gravity, magnetic, and electromagnetic data. The application of signal processing techniques in geophysics has led to significant advances in various areas, such as earthquake detection and location, imaging of subsurface structures, and exploration of natural resources, among others.

Some of the problems that can be solved in the field of geophysical signal processing are:

Current research in geophysical signal processing is focused on developing advanced signal processing techniques that can handle large and complex geophysical datasets. Machine learning techniques, such as deep learning and neural networks, are being used to develop automatic signal processing algorithms that can improve the efficiency and accuracy of geophysical data processing. Other research areas include the development of 4D imaging techniques that can monitor changes in subsurface structures over time and the development of joint inversion techniques that can combine multiple geophysical datasets for improved imaging of subsurface structures.

The butter function in MATLAB and Octave designs an IIR filter of a specific order n. When you use the filtfilt function to nullify the non-linear phase shift, it applies the filter twice, which effectively doubles the order of the filter. Therefore, the effective order of the filter is 2n.

If you need to use a specific order n for your signal processing, you should still specify n as the desired filter order in the butter function. The filtfilt function will then apply the filter twice, resulting in an effective order of 2n. When reporting the filter order in your research paper, you should report the order n that you specified in the butter function. If you used filtfilt to nullify the non-linear phase shift, you should also mention that the effective order of the filter is 2n. This is important for other researchers to understand the characteristics of the filter that you used in your signal processing.

Yes.. It looks okay. It can be futher denoised but it is better not to do it. You should allow some noise to ML to learn under some practical noisy condition.

Krishnapriya G B

To label the poles and zeros with names and different colors in the pole-zero plot, you need to add the following code after creating the transfer function t1:

hold on;

p = pole(t1); % extract the poles of the transfer function

z = zero(t1); % extract the zeros of the transfer function pzplot(t1); % plot the pole-zero diagram

title('Pole-Zero Plot of Transfer Function');

xlabel('Real');

ylabel('Imaginary');

% Plot poles with red x

for i = 1:length(p) plot(real(p(i)),imag(p(i)),'rx','MarkerSize',12,'LineWidth',2); end

% Plot zeros with blue o

for i = 1:length(z) plot(real(z(i)),imag(z(i)),'bo','MarkerSize',12,'LineWidth',2); end

legend('Poles','Zeros','Location','best');

hold off;

Siddharth Kamila Searching internet databases of academic publications such as Google Scholar, IEEE Xplore, and ACM Digital Library for research articles on the topic of the plot between real and imaginary components of an analytic signal is one approach to locate them. Use keywords like "analytic signal," "real and imaginary components," "complex signal," and "Hilbert transform" to locate publications on your topic.

Another option is to search the websites of signal processing conferences and publications, such as the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), the IEEE Transactions on Signal Processing, and the Journal of the Acoustical Society of America (JASA).

You may also wish to contact academics or experts in your field who might direct you to pertinent articles.

It is worth mentioning that the study article you seek may not be current, as the notion of analytic signal is a well-established issue in signal processing that has been intensively explored for many years.

The frequency content of P-waves and S-waves is not essentially different, so for me the unique clear differentiator remains their respective propagation velocities. You probably need at least two detectors kilometres apart to detect the wave nature from its propagation velocity e.g. by performing correlation between the time data: I doubt you can assess it from a single seismometer, where all waves and propagation paths will have interfered (unless you know you have a single strong seismic source not too far away and a sensor with very low background noise so that you can at a glance separate the successive waves of a single event in the time record: the P-wave is the first one you will detect). If the seismic source is very distant, the P-Wave is diffracted by the boundary layers of the successive layers of the Earth core and these various paths interfere, making the reading far more difficult... Furthermore, if there is a sequence of seismic bursts, the P-waves of the most recent one will be interfering with the S-waves of the earlier ones!

In brief, I do not see any escape from the current signal processing techniques of the world-wide seismic networks, I cannot believe frequency features alone will ever be determining specifically the P-Wave without using a distributed network of sensors to get more knowledge on the source event...

But this is simply stated from my own antecedent knowledge because you ask for external advices, I would not like to discourage your current research!

In such cases, you will need to develop the code based on the algorithm. Or look at source code of MATLAB and modify the code to suit your purpose.

answer of Anders Buen seems appropriate

Even you want to develop both the transmitter and receiver in your simulation, still you only need to do your simulation in baseband to include all effects of the channel. Actually, as I know, none one is doing such simulation in RF band, because all channel effects can represented in baseband simulation, while simulation in RF band is much expensive but provides none extra value.

Let's have a quick review of some concept. In general an RF signal (also referred to as bandpass signal) is:

P(t)=I(t)*cos(w*t)+Q(t)*sin(w*t),

where w is the angular frequency of the RF carrier, I(t) and Q(t) are the in-phase and quadrature-phase components of P(t) which is a real signal. For convenience, usually a complex signal S(t) = I(t)+j*Q(t) is defined and referred to as the baseband signal of P(t). Correspondingly, the bandpass (RF) signal can be expressed as:

P(t) = Real{S(t)*exp(j*w*t)},

where Real{x} stands for the real component of x.

With these notations, when the system (including the transmitter + channel + receiver) is studied, we only need to investigate the baseband signal S(t) and its changes (due to the channel).

In short, what you can do is as follows.

(1) Generate a sequence of baseband QAM symbols (denoted by S(t)).

(2) Generate a sequence of complex Rayleigh fading coefficients, such as

h(t) = a(t)+j*b(t), where a(t) and b(t) are independent Gaussian random processes with proper power spectral density depending on how fast the channel changes.

(3) The received baseband signal is r(t) = s(t)*h(t)+n(t), where n(t) is the optional noise. Your receiver then further works on this baseband signal r(t).

If you really want to see the waveform of the received RF signal, simple take the real component of {r(t)*exp(j*w*t)}.

Hope these helpful to you. Good luck.

The reconstruction of a sub-sampled signal by ideal low-pass filtering is affected by aliasing error. We have written the attached article about the aliasing error and some bounds of this error.

I think that these papers will be of some use to you ... with my best wishes for you

Digital Filter Performance Based on Squared Error

Convergence Rate For Low-Pass Infinite Impulse Response Digital Filter

Often known datasets like COCO are used for testing because it's well standardized and balanced. I don't know what kind of research you are doing, but you can see popular datasets here: https://imerit.net/blog/22-free-image-datasets-for-computer-vision-all-pbm/

If this is not what you are looking for, then you can search on Roboflow or Kaggle.

Aparna Sathya Murthy thanks. I understand that Butterworth/Chebyshev cant be part of DL. My question is if DL can substitute these filters in any way.

Hi Bjorn,

You are right when you discuss the Harmonic frequencies. In FFT you are taking time domain data and converting it into frequency domain data. You are assuming the FFT time domain continuing indefinitely. If you separate the ECG signal into points (you will arrive to certain number of samples) and the FFT assumes that these samples are infinitely repeated. All FFT have a frequency resolution that is equal to the sampling frequency divided by the number of samples. You will have certain sampling rate/ number of samples which would determine your frequency resolution. When you are using pacemaker, you are examining (pacemaker analyzes) frequency components of physiological signals and detect concealed properties to identify physical anomalies. Please read this paper PMID: 32051675 which would explain spectrum analysis algorithms such as Fast Fourier Transform (FFT), which are crucial for pattern detection. You have also listed common abnormal (path-physio ECG recordings) in Table 1. Values from these conditions indicate that in most common cases of pacemaker usage, the sample values encountered by the system will hold a high probability of being repeated over the course of time, hence you are able to find their main harmonic frequency and set up pacing such that it is recognized and reversed to your normal harmonics e.g. (Table 1).

Thank you all for your consideratio and response. The ppg is extracted from smartwatch and has many artifact noises. The frequency of them is not known. I read some articles that used wavelet for removing these artifact.first hpf is applied for removing the drift and then using wavelet.Recently I found a paper that use this method for EEG signal. Could it be possible to use that threshold for my ppg signal?

Dear Mihir Malladi ,

Data is the information we want to transmit. Signal is the waveform format used to send data over channel. Acts as carrier which carries data or payload.

https://www.rfwireless-world.com/Terminology/Difference-between-Data-and-Signal.html#:~:text=Data%20is%20the%20information%20we,to%20send%20data%20over%20channel.&text=Acts%20as%20carrier%20which%20carries%20data%20or%20payload.

Regards,

Shafagat

Have you considered the wfdb-python library (https://github.com/MIT-LCP/wfdb-python)? It has QRS evaluation support. It is a pure Python library. Hence it runs on all platforms without any need for compilation. In particular, in the wfdb.processing library, there is a function wfdb.processing.compare_annotations, which can be used to compare the reference annotations against a test annotation generated by your detector.

The demo notebook provided by them describes how to use it: https://nbviewer.org/github/MIT-LCP/wfdb-python/blob/main/demo.ipynb

In particular, you provide

- reference annotations [the list of indices at which R peaks have been marked]

- test annotations [the list of indices at which your QRS detector has detected R peaks]

- the window width where you allow a matching annotation to be found. Thus there may be a difference of a few samples/indices in the reference and your algorithm's annotations

- the ECG signal on which the annotations were done.

The function then identifies the true positives, false positives, and false negatives. It reports the sensitivity and positive predictivity of your detector. You can also access the list of indices for false positives and false negatives for further analysis.

Following is the rough sample code:

comparitor = wfdb.processing.compare_annotations(ref_sample=ref_sample,

test_sample=test_sample,

window_width=window_width,

signal=signal)

# print the results

comparitor.print_summary()

if comparitor.fp > 0:

print(f'False positives: {comparitor.unmatched_test_sample}')

if comparitor.fn > 0:

print(f'False negatives: {comparitor.unmatched_ref_sample}')

Dear Tao Li,

Your observation is very apt. Most of the time the academic articles may not contain adequate details for clear understanding. In such cases you should make you of cited articles to go down to the fundamental/original research works for better clarity. Books are better source for understanding how the methods and formulas are derived from scratch.

after reading an article baesd on quantum image encryption I think these two chaotic sequences are used for a key generation, not for encryption.