Sketch2Fashion: Generating clothing visualization from sketches

The field of unsupervised image-to-image translation in computer vision has undergone several developments giving rise to models that produce high-quality images while overcoming the one-to-one mapping used in earlier models. By leveraging the ability of these models, we undertake a project that aims to simplify the process of fashion design while preserving the creativity that is critical to the process by transforming sketches of fashion designs to final outfits, complete with textures and patterns. Our project experiments with three edge detection algorithms and tests out three models with different architectures. We perform qualitative as well as quantitative analysis through a human perceptual study to note possible advantages and disadvantages of the three models as they relate to the goals of our project. CS230: Deep Learning, Winter 2018, Stanford University, CA. (LateX template borrowed from NIPS 2017.)