The task is to predict the labels of images from the given categories of Buildings,Cars,Flowers,Faces,Shoes. We here train the dataset with differnt classifiers and compare the results.We try Neural Networks with backpropagation, Logistic Regression,Support Vector Machines. We further take our approach to apply DeepLearning methods and compare the results with our previous approaches. 1)Sparse Autoencoders 2)SelfTaught Learning 3)Deep Networks 4)Stacked Autoencoders 5)Linear Decoders 6)Convolution Neural Networks
The dataset consists of two parts: 'Training' folder contains training data: consists of Folder 'images' containing Training images Features file 'feature_vectors.txt': consisting of feature vectors for all the training images. Each row represents an image (instance) with the first column being image name, remaining 240 columns being the feature-vector of the image. The labels 'labels.txt': consists of Image name and the corresponding label/class (from {Buildings(1), Cars(2), Faces(3), Flowers(4), Shoes(5)}). 'Test' folder contains validation data: consisting of Folder 'images' containing test images