I really appreciated your two videos about Fourier Transform. You actually simplified almost everything and make it comprehensible. Not only that but also implementing everything from scratch using python with a detailed description. Thank you.
@@DigitalSreeni Yes, truly amazing! I am a biomedical engineering masters student and unfortunately my undergrad program did not require any coding besides just an introductory matlab course. This has proved to be a major disadvantage for me in my masters program because I am required to do a lot more coding for my masters courses. I am currently taking a Image Processing class that revolves around coming up with my own algorithms and I am not able to simply call pre-made modules to finish the projects required for the course. Do you have any resources available or know where I can look to get better at generating algorithms as opposed to simply calling the modules? You're videos at least provide an excellent stepping stone and they are easy to understand and comprehend!
At 9:30, could you please explain the purpose of “np.ogrid”, I’ve tried playing with it and it seems to assign a column vector to the first entry and a row vector to the second. How does the element selection work when you pass them into the function (circle), does it take every combination of x, y? Cause I tried meshgrid and it didn’t work :(
np.ogrid just creates a grid, like you realized it creates vectors for x and y. This is like creating a bunch of coordinates for your image space. Now that you have the coordinates, you can define any function using these x and y coordinates. A circle centered at (h, k)can be defined as (x - h) ** 2 + (y - k) ** 2 = r**2 So we can generate a circle inside your image using the x,y grid we just generated and with center being the central pixel location of the image.
Hello, if I have an image with 11x11 pixels and In the center of the image is a square of 5x5 pixels. The gray level of the background is 0 and the gray level of the square is 50. How can I compute the result of the magnitude of edges given by the compass operator for this image taking into account that the image is not noisy? I have the code but I don't know how to apply the math on paper...
Which one gives better results spatial domain or frequency domain? When to choose what domain, Could you please suggest something? Thanks for uploading such amazing content.
Beautiful demonstration. I have a question regarding your courses. Is there any video about image alignment, perhaps, using Gauss-Newton algorithm? Noone is using this algorithm for images, and probably you've worked with this for alignment or something else.
What do you mean by laser images, do you mean LSM? If so, deep learning may be the best approach. If you are looking for conventional approach then try Non-local means. Please watch my videos on denoising topic.
@@DigitalSreeni i have tried every filter that you have posted.One video suggests a point spread function and wiener filter which uses this works the best for me.Thanks for the tip
How to use the masked array in deep learning… I have set of images that has circular in center which doesn’t have any value(Nan). Now I want to use these images in deep learning. What should I do ?
Hello thaank u for this useful informations but if we wanna apply this fourier transform to a set of images , other way i have an image dataset and i want to apply on them this fourier transform how can i do it ?
This video is not about DFT and coding it from scratch, it is about using DFT as a tool to perform image filtering. I am sure you will find code on Github for DFT, if that is what you're looking for.
@@DigitalSreeni I want to ask you mister. If I use it for handling the imbalance data with combine using GANs. Is that possible? dataset -->> FFT ---> GANs ---> classification. Thank for your attention
![BLACK BAG - Official Trailer [HD] - Only in Theaters March 14](http://i.ytimg.com/vi/Du0Xp8WX_7I/mqdefault.jpg)
I really appreciated your two videos about Fourier Transform. You actually simplified almost everything and make it comprehensible. Not only that but also implementing everything from scratch using python with a detailed description. Thank you.
Glad it was helpful!
Better than my professor, wish you were teaching my course. Thank you!
Happy to help!
Oh man I thought this is very difficult. You made it so simple. Respect. Thank you very much. Subscribed.
oH MAN.!!! you are so briliant!! Keep going with that courses!! We love you so much!
Thank you :)
Very helpful and illustrative!
love your all videos sir, keep doing this nice work
Thanks a ton
Great videos, they are really well done and are so insanely helpful!
Happy to hear that!
@@DigitalSreeni Yes, truly amazing! I am a biomedical engineering masters student and unfortunately my undergrad program did not require any coding besides just an introductory matlab course. This has proved to be a major disadvantage for me in my masters program because I am required to do a lot more coding for my masters courses. I am currently taking a Image Processing class that revolves around coming up with my own algorithms and I am not able to simply call pre-made modules to finish the projects required for the course. Do you have any resources available or know where I can look to get better at generating algorithms as opposed to simply calling the modules?
You're videos at least provide an excellent stepping stone and they are easy to understand and comprehend!
At 9:30, could you please explain the purpose of “np.ogrid”, I’ve tried playing with it and it seems to assign a column vector to the first entry and a row vector to the second. How does the element selection work when you pass them into the function (circle), does it take every combination of x, y? Cause I tried meshgrid and it didn’t work :(
np.ogrid just creates a grid, like you realized it creates vectors for x and y. This is like creating a bunch of coordinates for your image space. Now that you have the coordinates, you can define any function using these x and y coordinates. A circle centered at (h, k)can be defined as (x - h) ** 2 + (y - k) ** 2 = r**2
So we can generate a circle inside your image using the x,y grid we just generated and with center being the central pixel location of the image.
Hello, if I have an image with 11x11 pixels and In the center of the image is a square of 5x5 pixels. The gray level of the background is 0 and the gray level of the square is 50. How can I compute the result of the magnitude of edges given by the compass operator for this image taking into account that the image is not noisy? I have the code but I don't know how to apply the math on paper...
Please watch my videos on convolution, e.g. video 95. Compass operators, like other operators, use kernels for image convolution.
Which one gives better results spatial domain or frequency domain? When to choose what domain, Could you please suggest something? Thanks for uploading such amazing content.
Beautiful demonstration.
I have a question regarding your courses. Is there any video about image alignment, perhaps, using Gauss-Newton algorithm? Noone is using this algorithm for images, and probably you've worked with this for alignment or something else.
based on your knowledge what would be the best filter that can be used to denoise laser images without information loss?
What do you mean by laser images, do you mean LSM? If so, deep learning may be the best approach. If you are looking for conventional approach then try Non-local means. Please watch my videos on denoising topic.
@@DigitalSreeni i have tried every filter that you have posted.One video suggests a point spread function and wiener filter which uses this works the best for me.Thanks for the tip
How to use the masked array in deep learning… I have set of images that has circular in center which doesn’t have any value(Nan). Now I want to use these images in deep learning. What should I do ?
Hello thaank u for this useful informations but if we wanna apply this fourier transform to a set of images , other way i have an image dataset and i want to apply on them this fourier transform how can i do it ?
May be this video helps: ruclips.net/video/QxzxLVzNfbI/видео.html
Sir, why r=0 to HPF does not get back the original image?
Can we say, HPF removes the low frequency components from the image?
High pass filter, as the name suggests, only lets high frequency components and blocks low frequencies.
Why can't I find a code from scratch on dft, everyone uses libraries after explaining the code. Makes no sense~
This video is not about DFT and coding it from scratch, it is about using DFT as a tool to perform image filtering. I am sure you will find code on Github for DFT, if that is what you're looking for.
How to make gaussian mask ?
permisiion to learn sir
Granted!!!
@@DigitalSreeni I want to ask you mister.
If I use it for handling the imbalance data with combine using GANs.
Is that possible?
dataset -->> FFT ---> GANs ---> classification.
Thank for your attention