The dataset used is from Kaggle ( Originally Presented in AnalyticsVidya competition ) comprises of around 25K images. The Repository aims to classify them in classes using Deep Learning and Machine Learning

The features for SVM Model was extracted using HOG( Histogram of Oriented Gradient ) Features.

The model was trained to achieve 41% accuracy for predicting the classes of the images

The CNN was implemented for classification using Tensorflow and Keras. The model was trained with 13 layers to achieve the Prediction Accuracy of 83%

https://www.kaggle.com/puneet6060/intel-image-classification/download