Fashion Clothes Generation System using Deep Convolutional GAN

Abstract

Abstract: The fashion industry has evolved into one of the most powerful industries in the world as a result of modernization. Before the middle of the 19th century, almostall types of clothing were made specifically for each person, either at home or on demand from dressmakers or tailors. Technological advancements such as the development ofartificial fibers, and nylon, as well as new dyeing and fabric cuffing processes, have given designers more creative flexibility. Likewise, the fashion industry has emerged various buying options like e- commerce platforms these days rather than the traditional approach. Where, some websites use automatic pattern generation in place of the conventional method (clothing designs). However, these websites are not likely to make high-end apparels accurately, which is why we propose to generate new fashionable clothes and develop a web application that generates high-end fashion apparels based on the training dataset by taking input from the users (the number of images that need to be generated by the model) using GAN technology, and letting the user choose colors for the generated apparels. GAN, short for Generative AdversarialNetworks, is a type of deep learning model that is used for generating synthetic data that is similar to the original data. It is composed of two neural networks - the generator and the discriminator - that are trained simultaneously to create and evaluate the synthetic data.For the purpose of creating high-quality fashion images, we suggest using Deep Convolutional Generative Adversarial Networks (DC-GANs). A deep learning technique called DC- GANs using convolutional layers in an adversarial network to produce images of a particular type. The "color palette" feature is implemented using a basic Image Processing Technique such as object color translation. Once the object is segmented, its color can be modified using various color transformation techniques such as RGB to HSV conversion or color balance adjustment