FASHION CLASSIFICATION USING MACHINE LEARNING

ABSTRACT

This study mainly focused on the underlying principles of fashion classification using deep learning. Due to the big data heavy consumption of resources, we used MNIST fashion dataset grayscale image in this research. There is heavy limit in the number of epochs as we only tested our model using 10 epochs due to computer used for this development. The higher the epoch the better but, stop-list has to be spotted out. Embedding the model for end-user consumption was not explained in this research as this will require use of big data technologies such as apache spark and hadoop, and we also have to apply the client-side technologies such as mobile device technologies android and iOS devices then fit our deep learning model into it, wish will be too challenging for this research. The programming language used in this project is the python programming language. The system in its all entirety is designed to completely eliminate the problem of the existing system. Traditional means of classifying fashion wears is tedious, prone to human error and unnecessarily expensive. The system when built will start performing the classification of fashion wears, hence, allowing the business owners to focus on building of what may be of interest to her potential customers.