Internet clothing shopping still faces problems such as inappropriate size selection, outfit mismatch, and limited visualization of clothing display on particular users. Virtual try-on systems have been developed to address the above problems, however many of the existing solutions are limited to original e-commerce platforms and do not consider handling reliable validation to prevent mismatched outfit combinations. The present disclosure relates to a cross-platform AI-based virtual fitting system for displaying apparel purchased from multiple online shopping platforms in a single try-on session. The suggested system allows a user to upload a full-body photograph and to submit links to top wear, bottom wear, footwear or complete attire products that come from different platforms, if any. In order to produce meaningful and realistic results, a multi-level mismatch validation system is built for assuring category correctness, gender fusion, and logical clothing-type consistency before initiating the virtual fitting process. User characteristics such as gender and other body measurements are taken into consideration to improve garment fitting and sizing. The acceptable clothing items are overlaid on the user's body-photograph by using image processing techniques to produce a realistic outfit display. The obtained results are stored temporarily on the local device, to protect user privacy.
Arunkumar et al. (Sat,) studied this question.