- 93
- 84 205
Morteza
United States
Приєднався 14 бер 2022
Subscribe to become scientifically enlightened 📚💻∞
In this channel, I am going to talk about the projects and research that I have done and am doing during my masters, PhD and postdoc. The topics that I would be covering are deep learning, deep convolutional neural network and their application, how to code them, image processing, computer vision, hyper/multi spectral imaging, spectral reflectance recovery, human color vision, digital cameras and their characterization, and Matlab and Python coding in a simple way.
In this channel, I am going to talk about the projects and research that I have done and am doing during my masters, PhD and postdoc. The topics that I would be covering are deep learning, deep convolutional neural network and their application, how to code them, image processing, computer vision, hyper/multi spectral imaging, spectral reflectance recovery, human color vision, digital cameras and their characterization, and Matlab and Python coding in a simple way.
How to Write Your Own Custom Loss Function in TensorFlow
In this video, I will show you how you can write your own custom loss (error) function in a neural network in TensorFlow. To access the code, use the following link:
github.com/mortezmaali/Custom_Error_Function.git
github.com/mortezmaali/Custom_Error_Function.git
Переглядів: 82
Відео
Best Way to Classify Hyperspectral Images in Python
Переглядів 22114 днів тому
In this video, I will classify an example hyperspectral image using the maximum distance algorithm. This is the best way, using which you can classify your hyperspectral data as it enables you to extract the independent endmembers in your data first. I already applied this algorithm in MATLAB, but here I am using Python. To access the code, use the following link: github.com/mortezmaali/Maximum...
A Short Tutorial on PCA and NNMF in Python
Переглядів 119Місяць тому
In this video, I will talk about Principal Component Analysis (#PCA) and Non-Negative Matrix Factorization (#NNMF) in Python. I will also draw a quick comparison between them. To access the code in this video, use the following link: github.com/mortezmaali/PCA-NNMF-in-Python.git
The easiest way to install sklearn, skimage and Colour libraries in Anaconda Python
Переглядів 165Місяць тому
In this video, the easiest way to #install #scikit-learn (#sklearn), #scikit-image (#skimage) and #Colour libraries in #Anaconda #Python is shown. These popular libraries are mainly used for machine learning, image processing and color science computations and learning how to easily install them can come in handy, as presented herein.
Colorimetric Characterization of Multispectral Cameras
Переглядів 1122 місяці тому
Multispectral cameras capture images across multiple wavelengths of light, typically including visible and non-visible bands like UV and infrared. They usually have between 3 to 10 spectral bands, though some can have more depending on the application. In this video, I will show you how you can colorimetrically characterize them so they reproduce the true color of the objects. To access the cod...
A basic image classification using CNN in TensorFlow
Переглядів 1342 місяці тому
Image Classification is a fundamental task in vision recognition that aims to understand and categorize an image as a whole under a specific label. In this video, a basic image classification using CNN in TensorFlow is presented. To access the code, use the following link: github.com/mortezmaali/Image_Classification.git
How to Save a Trained Neural Network Model and Use it Later in TensorFlow?
Переглядів 1292 місяці тому
In this video, I will show you how you can save a trained neural network in #TensorFlow and use it later. This can be very helpful, especially when we want the same answer from our model upon each run. To access the codes, use the following link: github.com/mortezmaali/Saving-a-Neural-Network-in-TF.git
Best Way to Classify Hyperspectral Images in MATLAB.
Переглядів 2473 місяці тому
In this video, I introduce to you the maximum distance algorithm, which is an automatic way of specifying and extracting the endmembers of a hyperspectral image. These endmembers are then used to classify the hyperspectral images. This is one of the most accurate way of specifying and extracting the endmembers of a hyperspectral image. To access the code, use the following code: github.com/mort...
Hyperspectral Image Classification Using SVM in Python
Переглядів 4323 місяці тому
Hyperspectral Image Classification Using SVM in Python
A Step-by-Step Guide to Solving a Simple Classification Problem Using SVM
Переглядів 814 місяці тому
A Step-by-Step Guide to Solving a Simple Classification Problem Using SVM
A Step-by-Step Tutorial on Classifying Handwritten Digits Using CNN
Переглядів 1074 місяці тому
A Step-by-Step Tutorial on Classifying Handwritten Digits Using CNN
Hyperspectral Image Classification Using CNN in TensorFlow
Переглядів 1,2 тис.5 місяців тому
Hyperspectral Image Classification Using CNN in TensorFlow
The Easiest Way to Install PyTorch in Anaconda Python (on Both CPU and GPU)
Переглядів 2,9 тис.5 місяців тому
The Easiest Way to Install PyTorch in Anaconda Python (on Both CPU and GPU)
What Causes Color Blindness and Do Men and Women Have the Same Chance of Developing It?
Переглядів 726 місяців тому
What Causes Color Blindness and Do Men and Women Have the Same Chance of Developing It?
The Easiest Way to Install Pandas, Matplotlib and Rasterio Libraries in Anaconda Python
Переглядів 1806 місяців тому
The Easiest Way to Install Pandas, Matplotlib and Rasterio Libraries in Anaconda Python
A Basic Hyperspectral Image Classification in Python
Переглядів 1,5 тис.7 місяців тому
A Basic Hyperspectral Image Classification in Python
Low Light Image Enhancement Using a Deep Convolutional Neural Network
Переглядів 5257 місяців тому
Low Light Image Enhancement Using a Deep Convolutional Neural Network
Best Way to Visualize Hyperspectral Data in Python
Переглядів 2,6 тис.7 місяців тому
Best Way to Visualize Hyperspectral Data in Python
A Basic Hyperspectral Image Classification in MATLAB
Переглядів 2,2 тис.8 місяців тому
A Basic Hyperspectral Image Classification in MATLAB
Basic points I wish I knew when I first started coding in Python
Переглядів 1648 місяців тому
Basic points I wish I knew when I first started coding in Python
What is the best way to read and save ENVI images of hyperspectral data in Python?
Переглядів 7079 місяців тому
What is the best way to read and save ENVI images of hyperspectral data in Python?
A practical example on how to turn an ordinary digital camera into a hyperspectral camera
Переглядів 7929 місяців тому
A practical example on how to turn an ordinary digital camera into a hyperspectral camera
How to synthesize a video from a sequence of frames using MATLAB?
Переглядів 12110 місяців тому
How to synthesize a video from a sequence of frames using MATLAB?
A short tutorial on linear regression in Matlab 2
Переглядів 6010 місяців тому
A short tutorial on linear regression in Matlab 2
How to automatically edit the names of a large number of files in MATLAB?
Переглядів 4411 місяців тому
How to automatically edit the names of a large number of files in MATLAB?
A tutorial on image registration in Python
Переглядів 67311 місяців тому
A tutorial on image registration in Python
How to build 3D models of objects in MATLAB using uncalibrated cameras using SfM?
Переглядів 727Рік тому
How to build 3D models of objects in MATLAB using uncalibrated cameras using SfM?
How to handle ENVI images of hyperspectral data in Python?
Переглядів 1,5 тис.Рік тому
How to handle ENVI images of hyperspectral data in Python?
how can ChatGPT help you with your job applications?
Переглядів 61Рік тому
how can ChatGPT help you with your job applications?
The easiest way to install TensorFlow in anaconda python
Переглядів 22 тис.Рік тому
The easiest way to install TensorFlow in anaconda python
Maybe I'm wrong but I think that you are using the same data for training and predictions (both X_train and X come from data) and also you are using train_test_split but then X_test is not used anywhere
@@g.gonzalez4097 the data have the same source but they’ve gone through different transformations…
Thank you, it worked for me! If someone is working with eps format, the solution of converting to pdf also works very well
Very Useful and interesting, thank you.
Very good.
Nice video, thank you.
How can I make an accurate image record between RGB and Long Wave thermal images??
@@elixassasin9671 can you expand on it a bit? Do you mean you want to register them to each other?
Yes, the thing is that I have images taken from a visible camera (1920x1080) and a thermal camera (640 x 480, I'm responsible for taking pictures of the same scene with a similar FOV, but obviously the visible camera has more coverage and apart you can see a scroll, I've used LightGlue to try to record them to make a small dataset with RGB images and its equivalent thermal image but so far I haven't had the best results, so I'd like to know if there is any technique. recommended to make this record as good as possible.
@@elixassasin9671 Yes, there are techniques for that, and it will be actually a topic of one of my next videos. There will be a paper published about the same topic soon, but it has not yet. But what I can share here is you need to look at the problem as trying to register two images of two different modalities. First of all, you gotta resize the images to make them the same size and then you should use something like this: www.mathworks.com/help/images/ref/registration.metric.mattesmutualinformation.html Of course, you have to tweak the variables to make them fit your own work. Email me if you have more questions. Hope this helps!
@@Morteza_Sc Oh thank you very much! I will check the resource you passed me to see what I can do with that, I appreciate your answer!
Thanks you Sir, for sharing, can I send you an email? i think your work would help me alot in my master thesis
Thank you so much! You can email me at morteza.maali.y2@gmail.com.
can we use the same code if the data is in xml format?
send me an email at morteza.maali.y2@gmail.com to send you a revised code to do that!!
Thanks you Sir for sharing. Waiting for your unsupervised learning like clustering experiment.
@@chainproject-q4n I will make a video on that! Thank you so much for your reminder!!
I just noticed that the following link also shows how to use Kmeans for hyperspectral image classification, which is an unsupervised method: ua-cam.com/video/hNrlsADn2CQ/v-deo.html I will make another video soon as well that uses a different unsupervised model.
Do you have code for calculating the signal-to-noise ratio of hyperspectral imagery? if yes then please provide the link for the same or if you can explain the procedure for calculating SNR would also be helpful. I really appreciate any help you can provide.
@@surendrakumarsahu7920 hi, I’m sorry, I’m away now but I’ll get back to you as soon as I’m back!! I am on an emergency trip!!
@@Morteza_Sc Thank you so much for your response sir, whenever you are free please help me on this.
@@surendrakumarsahu7920 Hello there! Here is a code that can potentially help you: github.com/mortezmaali/Hyperspectral_Image_SNR_Computation.git
@@Morteza_Sc Thank you so much!
@@Morteza_Sc Sir is it possible that I can use code without using ENVI as I have the dataset in the format of L1R, L1GST, and L1T only and I am unable to convert it into .hdr ? Or you could provide a dataset for the code that you have provided Or if possible can you tell me how I can get .hdr file for input.
tnx. that was very useful.
hello sir , I have followed the steps in the video by selecting Python version 3.9.19 on my computer, but when I wanted to install Keras GPU and TensorFlow, an error occurred that said my Python version was pinned. please help with the problem I am facing
@@ProgramerPinggirJurang delete the python you’ve installed using control panel and follow exactly what I say in the video. Don’t select the latest python model, just select whatever the version anaconda is suggesting you itself….
Not for beginners . I don't like it
@@SomeshCodes stay tuned, I’m planning to make a more beginner friendly video soon!!
This model need dataset that contains both RGB and hyperspectral images ??? because i do not have hyperspectral images
@@AmanSharma-vf7vs yes, every training models need two sets of data, training and testing. There are datasets out there where you can find the training and testing sets of your need. Email me and I can send the link to you…
100000000000000000000000000000000 times thank you
Will you please provide me the dataset
The dataset is here: github.com/mortezmaali/Image-Colorization2.git
It will be better if you speed up them when the fly up. And everything else is actually good
wow, absolutely nothing works... previously i installed it for cpu and it worked... after getting myself a new gpu, NOTHING WORKS FOR GPU... its over days, cant get it work on jupyter.
I can only imagine what kind of classified hyper-spectral classification algos that governments have created to use for military use. Imagine combining hyperspectral imaging with synthetic aperture radar data!
Thank you so much sir you have solve my big problem I was about to pay money ❤
Nice video, Pytorch might be more relevant now, though.
OSError: [WinError 126] The specified module could not be found. Error loading "C:\Users avee\AppData\Roaming\Python\Python312\site-packages\torch\lib\fbgemm.dll" or one of its dependencies.
@@MrRaveelhussain please make sure you follow the steps!!
The method interpolates the spectral data instead of measuring it. Thus it is a mock-hyperspectral data conversion from RGB to a plausible hypercube. Plausible means a linear solution with internal consistency.
Thanks you , you help me to save a lot of time
that was super satisfying, you have now gained a sub.
even after following all the steps. in anaconda command prompt a lot of files are being downloaded. why is that so?
@@md.shahriarabidswapnil604 those are the necessary files. Here you just use the easiest way to download and install those files.
@@Morteza_Sc yepp. I continued with those and verified. In sha allah wont face any problem
it was very helpful. thank you vai
GOD BLESSSSSSSSS YOU THANK YOU VERY MUCH §§§§§
excatly what i was looking for thanks man..
mrc :)
I have this issue "OSError: [WinError 126] The specified module could not be found. Error loading "C:\ProgramData\anaconda3\envs\torch_cuda\Lib\site-packages\torch\lib\fbgemm.dll" or one of its dependencies."
@@rishatdilmurat8913 please make sure you follow the instructions step by step… you shouldn’t get any error!!
@@Morteza_Sc I followed what you did on the video; The only difference was the version of torch and cuda. If you can send your email, definitely I want to send the version and step.
@@rishatdilmurat8913 send an email to morteza.maali.y2@gmail.com
surely this is still lossy?
@@SwithinFeely All methods are indeed… you have to be cautious with choosing a good training data. The training data should be a good representative of the testing data. That way, the amount of loss goes down.
I got a query i don't have predefined classes or labels for my hyperspectral data what can i use to create classes can i use 1d cnn!?
@@onkarkolekar7911 if you don’t have predefined classes, just look at my newest video which I just uploaded today. With neural network, you gotta have predefined classes…
i coudnt install tensor and keras
@@Abdurrahmanzangina please follow the instructions as listed there! Step by step!
hlo i didnt get the data file for working this python is there any other solution for this
The link to data is here: github.com/mortezmaali/True_color_reproduction_from_ENVI_Hyperspectral_Image_Python.git
Thank you for this well organized video. it has broken some barriers to my understanding of certain concepts. I also wanted to know to obtain the reference spectral data
@@ifeanyivincent370 thank you so much!! Please send me an email with data request and I send the data to you as soon as I can…
you beautiful beautiful man
Plz make video on Multimodal hyperspectral image classification using matlab
Thank you for making this great video.
the only tutorial which will really work in this year, thank you
Thanks for the video. This is the error I got (pins seem to be involved in the conflict. Currently pinned specs: -python 3.11.* ( labeled as pin-1))
@@goldgoddy1167 please delete and reinstall the python. And make sure you follow every steps the way I show in the video. You should not face any error. If you did, it means, somewhere you have done something wrong.
@@Morteza_Sc do i have to install the older version of python separately? or it will automatically install while creating environment?
@@realracing3specter295 just choose the option it has automatically chosen for you!
it looks like keras is now incorporated in the tensorflow package so you only need tensorflow
Hew. I did it. Thank you!
thanx for the vd, how about using CNN with Vit Transformer letting CNN extract the local and Vit extract the global features is there any vd for that?
Unfortunately there are not so many videos on the topic of deep learning application to hyperspectral image processing. I’ll look into it and let you know if I come across anything.
@@Morteza_Sc Okay thank you so much it's really appreciated
This was very helpful. Thank you
Hello ..How do i extract parameters from a 3D dense cloud??
To extract parameters from a 3D dense cloud created using Structure-from-Motion (SfM), you need to follow a few steps that involve analyzing the reconstructed 3D points and the associated camera parameters. Here are the key steps and some MATLAB code snippets to help you extract useful parameters: Load the Point Cloud Data: Ensure the xyzPoints variable contains the 3D coordinates of the points in the dense cloud. Visualize the Point Cloud: Use pcshow or other visualization functions to visualize the point cloud and understand its structure. Email me if you have more questions!
Thank you so much.
Great post! Do you have an example of ENVI HSI files to use with your code? Thanks
Thank you so much! This example uses ENVI images.
@@Morteza_Sc I meant - can you please post a link to example ENVI image? Repo didn't contain one, and I'm struggling to find example online. Thanks
@@georgiylevchuk8254 I see! I’m away now and don’t have access to the image that I used. I’ll be back in a few days and upload an ENVI image to the repository. I’ll let you know in a few days.
@@Morteza_Sc thanks!
@@georgiylevchuk8254 The data is now included: github.com/mortezmaali/True_color_reproduction_from_ENVI_Hyperspectral_Image_Python.git
Another great video, thanks Morteza. Can you add the hyper spectral image into the GitHub repo? Or point me to where I can find similar hyper spectral images?
Thank you, Morteza. Is there a paper that we can refer to?
Of course!! Unfortunately there’s no paper on this.