Tutorial 74 - What are Gabor filters and how to use them to generate features for machine learning?

For image processing and computer vision, Gabor filters are generally used in texture analysis, edge detection, feature extraction, etc. Gabor filters are special classes of bandpass filters, i.e., they allow a certain ‘band’ of frequencies and reject the others.
Code associated with these tutorials can be downloaded from here: github.com/bns...

[@tanaydeo8085]

Nobody could have explained it better !! Such a clear, simplistic yet robust video. Keep up with this fantastic work !

[@ZEISS_arivis]

Thank you :)

[@JoRoCaRa]

thanks! working on my thesis. you got yourself a subscriber! simple, to the point, well explained

[@ZEISS_arivis]

Thanks for the sub!

[@mamoon72]

i wish you were my teacher i would have been a scientist today really loved the way you make things not just theory but something which is learned to be applied

[@ZEISS_arivis]

Thanks for the comment. I hope you are learning from these videos.

[@deepeshmehta6107]

Your content is just awesome!! Gr8 explanation! Covering math, code and theory all in one complete package.

[@ZEISS_arivis]

Glad you like them!

[@lngl2786]

Finally got meaningful video. Thanks

[@RakeshKumar-kd1jh]

Thanks for explaining this concept in easiest way. Thanks a lot.

[@ZEISS_arivis]

Most welcome!

[@irishRocker1]

Really good tutorials, explained very nicely. I am definitely interested in learning more about the machine learning aspect of this. I am an electronic engineer for 10yrs but new to ML I was planning to do a project where some image processing and possibly machine learning would be used to improve results. It seems like this might suit. I had heard about a Joint bilateral filter that removes noise but preserves edges. Lot of research to do still!

[@KirillKulikovIkarus]

This was amazing! Very well explained! Thank you!

[@ZEISS_arivis]

Glad it was helpful!

[@ausialfrai3335]

Thank you Sreeni, yes I am very excited about Gabor filter and waiting for the next video :)

[@ZEISS_arivis]

Next video about Gabor is ready :) Very soon you will see videos about using Gabor and other filters for image segmentation. Stay tuned!

[@saigaak]

Thank you very much!!! You really helped me understand this. Thanks for great tutorial!!

[@johnmosh4772]

nice explanation

[@cedricvumisa7416]

thank you very much sir! good work

[@ZEISS_arivis]

Most welcome!

[@lksmac1595]

Very good knowledge!

[@ZEISS_arivis]

Glad it was helpful!

[@chenzakaim3]

amazing! thanks for the video

[@clcx6938]

thanks for your video

[@tilkesh]

Thank you very much

[@dr-anitag.khandizod9598]

very nice explanation .....Thank you sir

[@ZEISS_arivis]

Keep watching

[@sayandey1478]

Very coherent, thanks

[@ZEISS_arivis]

Glad it was helpful!

[@godexolrv4906]

nice explanation sir

[@ZEISS_arivis]

Keep watching

[@epampoefmkfkefpeao4291]

I was looking into how fingerprint scanners work today, and found many articles saying that the scanners use gabor filter for better image. Spent like the whole day looking what a gabor filter is... Was initially hoping for some program with UI... nope. Tried to run the python code. Didnt work either. So i ended up downloading some matlab app for applying gabor filter. A question arose after i did that, if the fingerprint scanners do actually use the gabor filter, do they apply it to small portions of the fingerprint at different angles, and then combine the whole image piece by piece ?

[@ZEISS_arivis]

I am sorry, I do not know how fingerprint scanners work. They may use gabor as this is a great filter to quantify texture.

[@YousefRoshdy]

Amazing, thank you so much Could please advise which mathematics course good for machinelearninig?

[@ZEISS_arivis]

You probably need statistics if you are interested in becoming a better data scientist. Just search for 'statistics for data scientist' on amazon and pick the book with maximum positive reviews. You may even find useful videos on KZbin, in case you learn via videos.

[@tahirarshad5801]

Hello hope you doing well. I am working on one project image classification through vision Transformer. I have question related Gabor Filter. when we apply 3D Gabor filter on image is there any gabor weights like(Gaussion Initialization weights) Generate or Not? If it is weights generate so how we can use this weights? Mean to say I want to use 3D Gabor filter weights to initialize hidden layer not with Gaussion initialize weights. Thanks

[@sohailali5741]

Thanks for great tutorial. I have road surface image in which I want to detect crack, but color of crack and pavement is almost similar and hard to distinguish. Can you please please let me know that will gabor filter help to enhance image? Particularly, can it help to differentiate crack and normal pavement? Your response will be highly appreciate. Thank you.

[@ZEISS_arivis]

I sure think so as crack is probably not as wide as the pavement. By applying Gabor filters of varying gamma you may find the right setting appropriate for crack detection. You just need to try.

[@robotcactus_mk1882]

Where is the video that goes over gabor filter selection?

[@ZEISS_arivis]

kzbin.info/www/bejne/r6jcoKCFrr2jrsk

[@ahmadirtisam9593]

Can we detect any Object using this Algorithm ?

[@ZEISS_arivis]

Gabor helps in extracting features. You need to use these features as input to a machine learning algorithm for object detection.

[@ilunasaa3740]

what is ktype ?

[@ZEISS_arivis]

ktype refers to the data format type for the kernel. In this example I used cv2.CV_32F which refers to float32.

[@clcx6938]

thanks for your great video.I find another formula for 2D gabor filter here inc.ucsd.edu/mplab/75/media//gabor.pdf, I'm confused that does the formula in video equivalent to the formula(30) in the doc? and how does they transform to each other?