Tutorial 83 - Image classification using traditional machine learning

Рет қаралды 15,128

Күн бұрын

This video provides an introduction to the process of generating features and using traditional machine learning (e.g. Random Forest, SVM) for image classification.
Image classification refers to labeling an entire image to a specific class (e.g. Cat, dog, barn, etc.). This approach allows you to use a handful of images to train an ML model to sort images based on texture and other features. The trained model can be used for future sorting of unlabeled images.
Code associated with these tutorials can be downloaded from here: github.com/bnsreenu/python_fo...

Пікірлер: 41

@15jorada 2 жыл бұрын

This guy saved my project single handedly. You are a gentleman and a scholar

@NoName4ify 2 жыл бұрын

Thanks a lot for this video!! Helped me a lot for my studies!! Great explanation!

@bhuvaneshwaripatil9269 2 жыл бұрын

Thank you sir, your videos helped in getting started with python coding. It will be helpful if you add contents for extracting texture features like LBP, BSIF

@maheralmasoodi8029 2 жыл бұрын

Great. Thanks 🙏 for efforts

@rainergog Жыл бұрын

This is really great! I am just doing an introductory image processing course at the University and was wondering how one extracts features from image filters and puts them into the tabular format that traditional machine learning algorithms require. Your video not only explained that matrix and vector reshaping very well, but makes all the steps in the process very understandable. Thank you, also for your other great videos!

@ZEISS_arivis Жыл бұрын

Glad it was helpful!

@MazenSOmran Жыл бұрын

Hello, first of all I want to thank you very much for the valuable content you provide and the beautiful way you present it. But here I have a question: Here, when working on classifying images at the level of the whole image, is there a way to estimate the importance of the feature as we did when working on image segmentation??

@RedRose-dn7sv Жыл бұрын

Thank you , it's very helpful 😀

@ZEISS_arivis Жыл бұрын

Glad it was helpful!

@yasaradikkubura6502 3 жыл бұрын

Many Thanks....Very informative.....

@ZEISS_arivis 3 жыл бұрын

You are welcome

@srivathsansanthanam639 11 ай бұрын

Huge Huge Thanks to Zeiss and Sreeni for coming up with such videos. Can someone in the comments or even Zeiss help me to determine Feature Importance in this case?.

@srivathsansanthanam639 11 ай бұрын

Issue sorted. Do the same feature importances. Output shud give the importances pixel wise for all the pixels in one image including filters. Say, u have a 5x15 image and applied 10 filters. Total pixels for single image is 5x15x10 = 750 pixels - so 750 pixels importance value we get. Divide this (rehsape this nparray) to 75x10. Now sum it along the axis containg 75 values. U end up with 10 values which are 10 importances for 10 filters

@Maxbario 3 жыл бұрын

Amazing content! Thank you, Screeni!

@ZEISS_arivis 3 жыл бұрын

My pleasure!

@guoshengsong8853 2 жыл бұрын

Thank you Sir， but I have one question when I use feature_extractor it always should "The parameter `image` must be a 2-dimensional array", I have no idea why it happens.

@SS-xx7jd 2 жыл бұрын

Thank you so much

@woonie3134 3 жыл бұрын

Very helpful my friend. How can u extract the text part of a label of a classified image when working with a large food meal dataset? I would like to classify food image and recommend recipes based on the label identified but I am not understanding how to do this. I WOULD kindly ask a tutorial on how to recommend the recipe. There is barely information on food related projects plz

@ZEISS_arivis 3 жыл бұрын

If you are referring to converting text images into text then you need OCR. There are many python libraries to do OCR.

@user-cq3mj4bt5w 2 жыл бұрын

Thank you very much sir please I wanted your help how can I used svm for image splicing classification in paytion

@arief1249 Жыл бұрын

Where to watch in your playlist that having same color conversion like these videos but using CNN?

@achillaode4395 11 ай бұрын

Thank you Sir， but I have one question when I just use 2 labels i got 0% acc, when i check i got 4 label. 2 label from train and 2 label from validation. and when I check using the confusion matrix I get 4x4 dimensions..

@futurefarming3394 3 жыл бұрын

Dear sir I will be very thankful to you if you upload some lectures on MRI segmentation if possible. Or any usable link (Like your videos)

@ZEISS_arivis 3 жыл бұрын

All the techniques from my videos apply for MRI images, including denoising. Please give it a try.

@santoshsmmarwar 2 жыл бұрын

Dear sir, thank u so much , your video is very nice. I m getting valueError : cannot reshape array of size 903168 into shape (8385, newaxis) Kindly provide the solution of it

@user-xe9em3ze7v Жыл бұрын

Why do you need to expand the dimensions before reshaping?

@rayhanuthama-5917 Жыл бұрын

how about to reshape image feature for CNN array, instead of 1,1,(array for value)? because i get this ambiguous data cardinality: ValueError: Data cardinality is ambiguous: x sizes: 1 y sizes: 160 (total train) Make sure all arrays contain the same number of samples.

@oindrubanerjee8587 Жыл бұрын

sir, I am getting "ValueError: cannot reshape array of size 245760 into shape (1644,newaxis)" this error, is during reshaping feture_image.how will I rectify it?

@gurucharank5491 Жыл бұрын

pls make a video on pyradiomics

@faribanamiranian456 Жыл бұрын

Hi Sreeni I watched your KZfaq videos and became interested. I wanted to know if I could read an article about image segmentation video (numbers 79 to 81 using traditional machine learning)? Thanks if you have an article send me

@ZEISS_arivis Жыл бұрын

Unfortunately we do not have many written articles on these topics. You may find some of our blogs useful. www.apeer.com/blog

@relationshipdestiny3355 Жыл бұрын

after this level, # Normalize pixel values to between 0 and 1 x_train, x_test = x_train / 255.0, x_test / 255.0, I had this error, please I need help with what to do x_train, x_test = x_train / 255.0, x_test / 255.0 TypeError: unsupported operand type(s) for /: 'list' and 'float'

@yahaisha 3 жыл бұрын

Dear sir, thank you so much, your videos really help my phd research.. i try this tutorial buat i'm having problem mentioned df is not defined ( ValueError: cannot reshape array of size 245760 into shape (190,newaxis).. )

@yahaisha 3 жыл бұрын

Dear Sir, problem solved ..tq:)

@ZEISS_arivis 3 жыл бұрын

I am glad you figured it out.

@bhavukthakur9501 2 жыл бұрын

Yahaisha, i am also getting same error can you tell how you solve the error?

@santoshsmmarwar 2 жыл бұрын

Hi dear , same problem i m facing of cantnot reshape array of size 903168 into shape ( 8385, newaxis) Kindly tell me how to solve this.

@shadatshohan4987 3 жыл бұрын

I'm having memory error to contain the pixels in the array...can you help?

@ZEISS_arivis 3 жыл бұрын

I am not sure how I can help other than suggesting to add memory or working with smaller data. You can try working on Google colab, even the free account will give you decent resources.