I want to classify iris images using simple machine learning algorithms like (SVM, LR, e.t.c). Is there an easy way to do so? I don't want to use AI or deep learning. Please suggest to me python or Matlab/Octave based procedures.
You can try with deeep learning framework (deepcognition.ai ) or peltarion AI tool
You have a nice tutorial for the image classification using Python's scikit-learn with SVM here: https://kapernikov.com/tutorial-image-classification-with-scikit-learn/
Thanks a lot for your answers....... I found the last suggestion pretty interesting.
SVM can be used but you would not have the luxury of convolutions.... SVM would just flatten the pixels into an array and find patterns... It may still be useful if the pictures are less complex else best to use CNN approach
I would highly recommend deep learning tool box in Matlab. This tool made designing a deep learning model for image classification like piece of cake.
For image classification using SVM, First, you need to apply feature descriptors like HOG, LBP, SIFT, etc. to extract the features from the images. thereafter you can apply SVM for the classification.
On the other, if you are using CNN, it will directly extract features from the images and classify them into the specified classes. In deep learning, you can use transfer learning for a small dataset and minimum computation requirements, it will better results compare to handcrafted features.
You can try deep convolutional neural network models for classification problems. Here are some nice tutorials for image classification:
https://machinelearningmastery.com/how-to-develop-a-convolutional-neural-network-to-classify-photos-of-dogs-and-cats/
https://machinelearningmastery.com/how-to-develop-a-cnn-from-scratch-for-cifar-10-photo-classification/
https://machinelearningmastery.com/how-to-develop-a-convolutional-neural-network-from-scratch-for-mnist-handwritten-digit-classification/
I Suggest you Deep learning tool box in Matlab if you comfortable with MATLAB. it is very easy and don't require any coding skills. Follow this procedure https://in.mathworks.com/help/deeplearning/ref/deepnetworkdesigner-app.html
As you do not want to use deep learning, I would suggest scikit-learn. It is an amazing library for Machine Learning in Python. You can apply most of the traditional machine learning techniques with few lines of codes.
https://scikit-learn.org/stable/
There awesome libs and methods. You need to know which of the fit to your data.
Here we have 10 of then: https://opensource.com/article/19/3/python-image-manipulation-tools
- Badges
- Science topic