Using this to help with a final project for an intro to ML class in university. Really helpful to have such a well organized set of explanations. Even with small errors mentioned in the comments this is a really helpful resource. Thanks for your work Rob!
Hi, just a correction, in the blurring section, to generate a more blurred image you need to change the size of the kernel too, not just divide for a large number, that's why you get a darker image instead of a blurred one. Thanks for the video.
compiler company should sponsor people like this, it will legit be the easiest way to get their compiler attention. Also thank you for making these videos and sharing your knowledge like this. Used it to unstuck myself for a uni project.
on 19:05 in my opinion you do not change the blur, only the pixel intensity by changing the kernel divider. if you want other blurs, you would have to adapt the kernel values themselves so that they are not equal anymore. right?
Hi Rob, just a quick novice's question. What are the advantages/disadvantages using Matplotlib vs OpenCV? Or in other words, what situation which library is better? I'm happy to find your tutorial, you aren't speaking like you have to win a "fastest talking competition" and you really share the principal informations leaving us to look into further details if necessary. (I'm not an Egnlish native)
Thanks for the feedback. The two libraries are very different and only really overlap in their ability to read and display images. OpenCV is a powerhouse for all things computer vision. Matplotlib can do plotting of data and some image processing.
ditect place and shape and size of upload second image in main image and after ditectin all thing second image upload in main image acording to ditection and save image i dont get solution please help me sir give me solution
For a university project, I must read a car registration with Python, get its characters. I don't know if you are interested in doing something with this. A camera that reads the car registration, collects the characters and can use those characters to compare it in a database.
@@robmulla thanks for answering. You are very good teaching I am from Colombia and the automatic translator works very well. I can read everything you say in the videos and really teach too well. Congratulations.
When we flatten out the matrix and plot it with pd.Series(img_mpl.flatten()).plot(kind='hist', bins=50, title='Distribution of Pixel Values'), because of each pixel is represented by 3 subpixels (RGB) in the matrix it kind of not made sense to me. Maybe labeling it as "Distribution of RGB Values in Pixels" might be a better approach if I am not missing something?
Really good presentation and content flow. I think THIS video would have been better without the kernel bits since it wasn't clear what the numerical values (of the kernels) were doing. A video focusing on kernels would actually be excellent all on it's own (with a deeper dive into them and what they're doing). Overall, still an excellent video though! Thanks!
Thanks Dan, that's great feedback. I'm glad you liked the overall flow of the video. Now that you mention it the kernel stuff at the end was a bit forced. I keep trying to remind myself that short and sweet videos are better than overloading with too much information. Thanks for the feedback and I'll keep that in mind for the next video!
The representation of topic was very clear, precise and nice. I wish to see such tutorial on color image processing with wavelet transform and quaternions color image processing. Please make such video if possible. Thank you for such nice video
Thanks Aka. I'm still working up towards more advanced machine learning videos but hope to get to them in the future. Hopefully this video is still helpful for anyone brand new to image data working on an image kaggle competition.
iyi günler, aklımda subway surfers demo oyunu yapmak fikri var ancak bu oyunu obez insanlar, evde hareket etmek isteyen insanlar ve çocuklara yönelik yapmak istiyorum. Burada amacım telefonun ön kamerasını kullanarak image processing kuralları çerçevesinde hareketleri algılayıp oyun karakterini engellerden kaçındırmak. Şöyleki sürekli koşar vaziyette olacağım karakter eğilmesi gerekiyorsa eğileceğim zıplaması gerekiyorsa zıplayacağım sağa kaçması gerekiyorsa sağa kaçacağım vb. 1 yıllık mobil deneyimim var ancak image processing hakknda lafta bilgim var. Nasıl bir yol izlemeliyim hangi konularda uzmanlaşmalıyım ve 3-4 ay içinde yapabilir miyim? Kendimi adamaya hazırım
Nice intro video. If you wanted future video suggestions, can you dive a bit deeper into how to manipulate those numpy arrays that contain image data? So things like np.transpose, np.where, using zip to extract pairs of coordinates, and so on. I know some of this isn't opencv perse, but that knowledge seems key to doing more complicated image processing, object detection, etc.
Thanks for the suggestions. Diving deeper into numpy expressions would be important when working with images in numpy format. Maybe in the future I could cover it.
thank you for your tutorial. What if I want to create for example animations based on the numpy array of an image. What next steps you suggest? Thanks!
Thanks for your tutorial. But I have a question. This is my code example: import numpy as np import cv2 import matplotlib.pyplot as plt img = cv2.imread('D:/CCU-2024/grasppose/p1.jpg',0) plt.figure() plt.subplot(1,2,1) plt.imshow(img) plt.show(). So, when I move the mouse on the image of plt.imshow(). I obtain the (x,y) and [a]. So what is (x,y) ? it's (u,v) right? and [a] is pixel value, right?
Thanks for your videos! I recently started watching your videos, and kaggle keeps failing to save my draft and I lose my notes, what can I do about this?
Excuse me Sir, please can I ask you a question: I just decided to pay for the OpenCV certification, its 1500 canadian dollars, and I m pretty convinced that it is a great investment for my future career in CV. Please can you give me your opinion? It will be a great help Thank you in advance :)
Thanks for watching. It should be really easy to do manually using numpy.flip or you could flip using the built in cv2. Here is the numpy documentation you would just need to select the correct axis: numpy.org/doc/stable/reference/generated/numpy.flip.html
Using this to help with a final project for an intro to ML class in university. Really helpful to have such a well organized set of explanations. Even with small errors mentioned in the comments this is a really helpful resource. Thanks for your work Rob!
Love that you found the video helpful. You should share it with the rest of your class! And have them share…. 😁
Hi, just a correction, in the blurring section, to generate a more blurred image you need to change the size of the kernel too, not just divide for a large number, that's why you get a darker image instead of a blurred one.
Thanks for the video.
Excellent point. Thanks for pointing that out.
Change the size to what? How does it have to match the number you divide by?
@@bofa722just change the kernel size it will averages more pixels thus more blurring
compiler company should sponsor people like this, it will legit be the easiest way to get their compiler attention. Also thank you for making these videos and sharing your knowledge like this. Used it to unstuck myself for a uni project.
Excellent introductory video, well done. I'm highly motivated to go off and explore now.
“Curiosity driven data science” is the mission statement of my channel. This comment makes me so happy because this video inspired you to be curious!
Thanks again Rob... Always very high quality content and lessons!
Thanks Filippo!
19:05 For changing the blur amount, re-adjust the kernel size. Higher the kernel size, blurrier the image gets
this video is very important to me thank you mr.rob mulla from sri lanka
@02:44: one second of OCD. Thank you!!!
Great video! Looking forward to watching more videos like this one. Wish you all the best ❤️
Thanks so much Kha! I'm glad you found the video helpful.
on 19:05 in my opinion you do not change the blur, only the pixel intensity by changing the kernel divider. if you want other blurs, you would have to adapt the kernel values themselves so that they are not equal anymore. right?
Great video as always!!
I apprecaite that a lot Akshat!
Hi Rob, just a quick novice's question. What are the advantages/disadvantages using Matplotlib vs OpenCV? Or in other words, what situation which library is better?
I'm happy to find your tutorial, you aren't speaking like you have to win a "fastest talking competition" and you really share the principal informations leaving us to look into further details if necessary. (I'm not an Egnlish native)
Thanks for the feedback. The two libraries are very different and only really overlap in their ability to read and display images. OpenCV is a powerhouse for all things computer vision. Matplotlib can do plotting of data and some image processing.
ditect place and shape and size of upload second image in main image and after ditectin all thing second image upload in main image acording to ditection and save image
i dont get solution please help me sir
give me solution
For a university project, I must read a car registration with Python, get its characters. I don't know if you are interested in doing something with this. A camera that reads the car registration, collects the characters and can use those characters to compare it in a database.
Sounds like a cool project. I’ve been asked this type of question a lot so maybe I’ll try to make a video about it. Good luck!
@@robmulla thanks for answering. You are very good teaching I am from Colombia and the automatic translator works very well. I can read everything you say in the videos and really teach too well. Congratulations.
13:14 it's not just grey scaled, it's also inverted
Hi Rob, I want to thank you for sharing your videos. I' ve found them very useful to start my way on data science. Cheers from Argentina!.
Mr Mulla, you are not a genius, you are a god ! Thanks for the inspiration !
When we flatten out the matrix and plot it with pd.Series(img_mpl.flatten()).plot(kind='hist', bins=50, title='Distribution of Pixel Values'), because of each pixel is represented by 3 subpixels (RGB) in the matrix it kind of not made sense to me.
Maybe labeling it as "Distribution of RGB Values in Pixels" might be a better approach if I am not missing something?
Really good presentation and content flow. I think THIS video would have been better without the kernel bits since it wasn't clear what the numerical values (of the kernels) were doing. A video focusing on kernels would actually be excellent all on it's own (with a deeper dive into them and what they're doing). Overall, still an excellent video though! Thanks!
Thanks Dan, that's great feedback. I'm glad you liked the overall flow of the video. Now that you mention it the kernel stuff at the end was a bit forced. I keep trying to remind myself that short and sweet videos are better than overloading with too much information. Thanks for the feedback and I'll keep that in mind for the next video!
thanks, great video ! just a correction though - when you make the image smaller by x in each axis you actually make the whole image x^2 times smaller
Oh, great point! Thanks for pointing that out.
digital art should be made like this
The representation of topic was very clear, precise and nice. I wish to see such tutorial on color image processing with wavelet transform and quaternions color image processing. Please make such video if possible. Thank you for such nice video
# Dear Rob, thank you for sharing, you knowledge, it would be very helpful if we could apply given tutorials, on some kaggle competitions
Thanks Aka. I'm still working up towards more advanced machine learning videos but hope to get to them in the future. Hopefully this video is still helpful for anyone brand new to image data working on an image kaggle competition.
hello sir, can you make a video for image processing with opencv using python
while working the same in Jupyter notebook , it gives list out of index error while reading the images of my folder. why is that so?
hi please i need help with a topic speaker identification in python. how convert dataset to array two dimension
Hello teacher. I have a question.
Why the numpy array of PNG image have range 0~1, JPG have range from 0~255?
iyi günler, aklımda subway surfers demo oyunu yapmak fikri var ancak bu oyunu obez insanlar, evde hareket etmek isteyen insanlar ve çocuklara yönelik yapmak istiyorum. Burada amacım telefonun ön kamerasını kullanarak image processing kuralları çerçevesinde hareketleri algılayıp oyun karakterini engellerden kaçındırmak. Şöyleki sürekli koşar vaziyette olacağım karakter eğilmesi gerekiyorsa eğileceğim zıplaması gerekiyorsa zıplayacağım sağa kaçması gerekiyorsa sağa kaçacağım vb. 1 yıllık mobil deneyimim var ancak image processing hakknda lafta bilgim var. Nasıl bir yol izlemeliyim hangi konularda uzmanlaşmalıyım ve 3-4 ay içinde yapabilir miyim? Kendimi adamaya hazırım
Nice intro video. If you wanted future video suggestions, can you dive a bit deeper into how to manipulate those numpy arrays that contain image data? So things like np.transpose, np.where, using zip to extract pairs of coordinates, and so on. I know some of this isn't opencv perse, but that knowledge seems key to doing more complicated image processing, object detection, etc.
Thanks for the suggestions. Diving deeper into numpy expressions would be important when working with images in numpy format. Maybe in the future I could cover it.
i was thinking in using the image numpy array to create animations programmatically manipulating the values. Sounds coherent?
thank you for your tutorial. What if I want to create for example animations based on the numpy array of an image. What next steps you suggest? Thanks!
Hi Rob thanks for the tutorial! Can you make one on analyzing trends with a large volume of images, common colors, materials, objects etc ?
Awesome video, thanks man
13:22 Gray scale but also negative
Thanks for your tutorial. But I have a question. This is my code example: import numpy as np
import cv2
import matplotlib.pyplot as plt
img = cv2.imread('D:/CCU-2024/grasppose/p1.jpg',0)
plt.figure()
plt.subplot(1,2,1)
plt.imshow(img)
plt.show(). So, when I move the mouse on the image of plt.imshow(). I obtain the (x,y) and [a]. So what is (x,y) ? it's (u,v) right? and [a] is pixel value, right?
please do tutorial of this with tar file. am trying to open a custom dataset with tar file. having trouble opening
I think the chanel images are inversed like a negative
Oh yea? Do you mean with RGB?
I mean like the negative on a film photography
hii can i do the same in jupyter notebook?
as i work on that
Which interface are you using? Is it google colab pro
This is a kaggle notebook. Check out my tutorial on jupyter notebooks and I explain my setup.
How can I link this model with interface
I need your help please reply to me In the fastest time
Plz use geotiff images for GIS use as well in next lecture.
Thanks for your videos!
I recently started watching your videos, and kaggle keeps failing to save my draft and I lose my notes, what can I do about this?
Excuse me Sir, please can I ask you a question: I just decided to pay for the OpenCV certification, its 1500 canadian dollars, and I m pretty convinced that it is a great investment for my future career in CV.
Please can you give me your opinion? It will be a great help
Thank you in advance :)
4:23 was that pun intended?😆
what is that platform u using?
Linux.
Wonderful tutorial! But upscaling width and height 10 times simultaneously doesn't make the image 10 times bigger, but 100 times bigger
Kindly make a video on a cardiomyocyte Data Analysis using Python.
Dear data 👨🔬
Let me know about initial image data for face and accidental face data to find the difference in uaing open cv and matplotlib🎉please sir
Why use OpenCV instead of Scikit-Image?
Thanks for your video 😊
Thanks for watching!
Is it work for svg format
Would be great if you could have explained how to flip an image with the usage of pixels
Thanks for watching. It should be really easy to do manually using numpy.flip or you could flip using the built in cv2. Here is the numpy documentation you would just need to select the correct axis: numpy.org/doc/stable/reference/generated/numpy.flip.html
Do one with detecting different colors within a pic
It looks like you are working in Linux. I have job interview. The source data come from TV. Can you teach me?
Yes I use Linux. My Chanel is my teaching. Hope you find it helpful.
sneaky suplots to subplots @6:49
Nice catch! Must have fixed it and edited that part out 😅
@@robmullayeah xd, but anyway thank you for your videos and Happy new year from the future! pd.Timestamp.now(tz='Etc/UTC+1')
I'm from Pakistan your best video for welcome
hi thanks for this video,this video has 5 likes.😃
Thank you too. Hope you liked it!
Thx
How do masks work?
Depends what type of mask we are talking about...
python is for fast prototyping, not for long lasting code.... new releases of whatever in python kills previous code.... what a shame!!!!
Gotta love heavy abstractions
BORRRRRRRRRINF WHHYYYYY
.
@@LOL_CYABOYwhat is bro yapping about