273 - What is Voronoi - explanation using python code

Code generated in the video can be downloaded from here:
github.com/bns...
A Voronoi diagram divides the plane into separate regions where ​each region contains exactly one generating point (seed) and​ every point in a given region is closer to its seed than to any other. ​The regions around the edge of the cluster of points extend out to infinity.
This video explains Voronoi using python.

[@dnakama2096]

Terrific! I have been the whole day looking for information of Voronoi and this video let me to understand it completely. Thanks for uploading this high quality content!

[@krishnanunnib4546]

Vornoi is also used in 3d softwares for making fractures in a geometry.

[@joelgraff6550]

Another great video, Screeni. There were a couple times where you were pointing out something on the video, but your mouse pointer wasn't visible. Although I could understand where you were talking about in the image, seeing a mouse pointer could help make this clearer. Also, do the lines in the Voronoi texture run perpendicular to an unseen line connecting the two vertices? It seems like the line is close to perpendicular, but it's difficult to be sure because I feel like there was a hint of optical illusion that made my mind a little confused.

[@DigitalSreeni]

Yes, I realized that too late. Need to be mindful for future videos.

[@joelgraff6550]

@@DigitalSreeni Congrats on hitting 3,000,000 views.

[@zhenli1630]

could you tell me that how to plot weighted voronoi by python, or can i set a value for the expanding speed from the scatter?

[@DrawMeAParadox]

Hello sir! First of all, thank you very much for what you do for us 🙏 I have a question: how can we extract random points from a plot into an array or a list?

[@rams8265]

Can you please share the link for the temp.jpg file. Thanks, very nice summary. Looking forward to the next one.

[@TuTu_Investing]

Hi Sreeni, thank you very much for another great video! I wonder where we could use this voronoi approach in a more powerful and specific way where a simple median filter followed by a thresholding, a watershed and find contour could not solve.

[@DigitalSreeni]

These are all the tools you can use to achieve a specific task. We still need watershed with Voronoi approach; it is just that Voronoi makes it easy to identify the 'domain' around an object that needs to be segmented.

[@ashwiniyadav464]

Hello sir.. Very interesting and informative video thank you very much Sir can u suggest latest segmentation method for medical image applications

[@user-nv2xy8dx5b]

Thank you sir for what you are doing... Sir, I have trained effiecientnetbo to classify images and then convert them to tensorflow lite I am trying to deploy on Raspberry pi 400 but the message appears when I try to load the model "could not open 'model path' s"

[@DigitalSreeni]

Looks like the path to your model is not correct. You may have to play with slashes and backslashes as things are different between Windows and Linux.

[@younesmatoqu2573]

Thank you very much for sharing your great knowledge with us I highly appreciate your effort, can you please discuses Mask R-CNN since it is important nowadays in most of the fields, waiting for this video from you cause you are the best in teaching!!!!

[@DigitalSreeni]

Coming up in a few weeks. Stay tuned.

[@omarfaruqehamim8205]

You said that each voronoi region contains exactly one point. But in the illustration at timestamp 2:09, I can see upper-right most (blue colored) and lower-right most (brown colored) regions do not contain a single point. I wanted to ask what is the reason behind this, and also is it possible to have less number of voronoi regions than the number of points?

[@PUBUDUCG]

Why this can not be done with simple thresholding ?

[@lalitsingh5150]

Thanks for another great Video.....

[@DigitalSreeni]

Thanks for watching!

[@lomacomfktoo200]

Hello Sir, thank you so much for always teaching us 🙏, as usual great video and content of quality 👍 this one came right in time😆 almost like u'r watching over us 👀😅 just asking .... the peak_local_max are not corresponding to the img... or it's me who's missing something?🤔 i was trying to plot the voronoi over the img when i noticed, i fliped and rotated the img to match the points. ------------------- vor=Voronoi(points) fig,ax = plt.subplots(1,1) ax.imshow(np.rot90(np.flip(blured_img, 0), k=3), cmap="binary_r") voronoi_plot_2d(vor, ax, line_colors="yellow") plt.xlim([0,256]), plt.ylim([0,256]) # plt.axis('off') plt.show()

[@DigitalSreeni]

I don't expect any image flipping during peak_local_max operation. Not sure what's going on with your images. Must be something else.

[@lomacomfktoo200]

​@@DigitalSreeni yeah i tried: for coords in points: blured_img[tuple(coords)]=0 plt.imshow(blured_img) plt.show() every point on spot....then i noticed: plt.plot(points[:,0], points[:,1], 'r.')# what i did plt.plot(points[:,1], points[:,0], 'r.')# what u did the problem was just that plt.plot and voronoi_plot_2d use a different axis order than plt.imshow. so if we want to plot one over the other we need to invert the points coordinates or rotate and flip the img. Thanks for your reply🙏

[@hrushikeshsahasrabuddhe5103]

Superb video!

[@simonemartino9425]

Is possible to compute it in 3D and visualize ?

[@nying3452]

Dear sir, this is a very wonderful video. I would like to use this tool. but I can't install this with pip and conda. could you guide me on how can I install this, sir?

[@tilkesh]

Thx

[@RAZZKIRAN]

thank you sir

[@DigitalSreeni]

Most welcome

[@naimamirobidova2092]

Thank you

[@DigitalSreeni]

You're welcome

[@safwanshaibani8928]

Sir, the code is not there on Github, Can you mentioned again ?

[@DigitalSreeni]

It will be uploaded tomorrow when I return back home from my business trip. Thanks for your patience.

[@aliyousif9319]

Thank you

[@DigitalSreeni]

You're welcome