I have basic knowledge of SVM, but now I am working with images. I have images in 5 folders, each folder, for example, has images for letters a, b, c, d, e. The folder 'a' has images of handwriting letters for 'a, folder 'b' has images of handwriting letters for 'b' and so on.
Now how can I use the images as my training data in SVM in Python.
as far i understood you want to train your svm to classify these images into the classes named a,b,c,d . For that you can use any of the good image processing techniques to extract features (such as HOG which is nicely implemented in opencv) from your image and then use these features , and the label as the input to your SVM training (the corresponding label for those would be the name of the folders i.e a,b,c,d) you can train your SVM using the features only and during the inference time , you can simply calculate the HOG feature of the image and feed it to your SVM and it will give you the desired output.
