Image Classification using sklearn

Using Machine learning algorithms to classify images under 3 categories

sklearn

sklearn or scikit learn is a library in Python with efficient tools for machine learning and statistical modelling. This project uses the SVM or Support Vector machine module under sklearn library to classify images under 1 of 3 categories

Support Vector Machine (SVM)

SVMs are supervised machine learning algorithms that are used for 2 group classifications (They can be used for more than 2 classes, by changing the kernel parameters). Data is classified when an optimal hyperplane is drawn, maximum distance away from the data points. All data points with minimum distance to the hyperplane are called Support Vectors. In this code, we have 3 classes to classify:

Cute Puppy
Red rose
Ice cream cone

Kernel

The popular kernels in sklearn are: linear, polynomial, radial basis function(rbf) and sigmoid. Kernel is used to map the distance between 2 points of the dataset. A score is assigned to the distance, where closer the points, higher the score. This score can be used for mapping points to more dimensions (kernel trick). RBF kernel is used in this code, as it gives good classification report results (compared to others). Best parameters are evaluated using GridSearchCV() function in sklearn.

Libraries used

Refer to requirements.txt to see which files and versions are necessary to run this code

References

https://scikit-learn.org/stable/modules/svm.html

https://towardsdatascience.com/multiclass-classification-with-support-vector-machines-svm-kernel-trick-kernel-functions-f9d5377d6f02

https://www.youtube.com/watch?v=dw96S_iFFbI&t=7346s