Classification of images using Convolutional Neural Network.
It is very easy to recognise the image for a human one, but for a computer it is very difficult to do that. Convolutions Neural Network allows us to analyse the image content. CNN can be implimented using the sklearn library, CNN model take images as input to model and after the classfication we get the result in predicted classes.
Keywords : Machine Learning, CNN, Image Processing, Matplotlib.
PreRequirements :
LIBRARIES : Sklearn, Keras, matplotlib.
IDE : spyder
Classfy the images and store them into respective folders, built the model using keras
Procedure to Exicute the Program:
1). Exctraction :
Dataset is Downloaded form Kaggle
2). Preporcessing
Images are proprocessed and no need to resize and if needed we can resize them using numpy.
3). Model Training
Run the model.fit statement
4). prediction
To predict the images content run the model.predict command with argument as a numpy matrix
As this is a Classification Problem we can calculate the accuracy of model from sklearn.metrics using accuacy Score.