Image-based Virtual Fitting Room
This addresses the problem of limited fashion item detection and poor style changes in virtual fitting rooms for e-commerce and fashion design, though it is incremental.
The paper tackled the challenge of virtual fitting rooms by proposing a method that combines Mask R-CNN for detecting fashion items and Neural Style Transfer for altering their style, achieving 68.72% mAP and 0.2% ASDR, outperforming baselines.
Virtual fitting room is a challenging task yet useful feature for e-commerce platforms and fashion designers. Existing works can only detect very few types of fashion items. Besides they did poorly in changing the texture and style of the selected fashion items. In this project, we propose a novel approach to address this problem. We firstly used Mask R-CNN to find the regions of different fashion items, and secondly used Neural Style Transfer to change the style of the selected fashion items. The dataset we used is composed of images from PaperDoll dataset and annotations provided by eBay's ModaNet. We trained 8 models and our best model massively outperformed baseline models both quantitatively and qualitatively, with 68.72% mAP, 0.2% ASDR.