Building a powerful Neural network that can classify Natural Scenes around the world
The goal is to classify images under 6 categories. {'buildings' -> 0, 'forest' -> 1, 'glacier' -> 2, 'mountain' -> 3, 'sea' -> 4, 'street' -> 5 }. The dataset is available for download at https://www.kaggle.com/puneet6060/intel-image-classification. I have implemented a CNN model and a DenseNet121 and ResNet50 model (pretrained on imagenet) for the classification tasks. The CNN model achieved a test accuracy of 90.62% whereas the DenseNet121 model achieved a test accuracy of 92.2% and the ResNet50 model achieved an accuracy of 95.3% on the test set.