Boosting the Accuracy of an Image Classifier - IT Assignment Help

Assignment Task

Abstract

Image classification is a supervised learning problem : define a set of target classes (objects to identify in images), and train a model to recognize them using labeled example photos. Early computer vision models relied on raw pixel data as the input to the model. However, raw pixel data alone doesn't provide a sufficiently stable representation to encompass the myriad variations of an object as captured in an image. Now convolutional neural networks (CNN) could be used to progressively extract higher- and higher-level representations of the image content. In this project we relied on improving more and more accuracy. So first of all, we have created an Image classification model with the cifar10 dataset, which is a standard dataset containing lots of images of 10 classes. These 10 classes are cats, dogs, airplanes, birds, ships, frogs, horses, automobiles, deers, and trucks. We created the classifier with the use of Keras library along with the cifar10 dataset and trained it to achieve an accuracy of around 81%. Then, we tried to improve the accuracy of that model using a technique called ‘boosting’. We will use the Adaboost algorithm to boost convolutional neural networks.

Further, In eagerness to achieve much more accuracy ,we tried to create another image classifier with the help of PyTorch library and better thing is that we achieved an unexpected 97?curacy. (PyTorch - it is an open source machine learning library which has varied applications in computer vision and natural language processing, primarily developed by FB’s AI Research lab). PyTorch makes it easy to load pre-trained models and build on them. Some of the most popular pre-trained models, like ResNet, AlexNet, and VGG, come from the ImageNet Challenge. These pre-trained models allow others to quickly obtain cutting-edge results in computer vision without needing such large amounts of computer power, patience, and time. We actually had great results with DenseNet and decided to use DenseNet161, which gave us very
good results relatively quickly. Next we built our classifier, trained our model and tested with a sample image. One of the interesting things is that we achieved 95% in the first 7 epochs only. Hence at last we reached our goal of achieving 97?curacy.

This IT Assignment has been solved by our IT experts at My Uni Paper. Our Assignment Writing Experts are efficient to provide a fresh solution to this question. We are serving more than 10000+ Students in Australia, UK & US by helping them to score HD in their academics. Our Experts are well trained to follow all marking rubrics & referencing style.