Welcome to JJEM: A Multi-Disicplinary Journal of JNNCE, Shimoga |
Girish Mantha, Praveen G Shet,, Rahul Athreya K M, Sathwik P Bhat, , Shamanth G P
The Virtual cloth Try-on is one of the biggest inventions took place in fashion industry which contributes to enhance user experience by allowing them to try out garments virtually without wearing it. The Virtual Try on Cloth is image-based technology to enhance the user experience on fashion-oriented e commerce websites, it will help customer’s satisfaction. To get perfect body fit cloth or exact cloth fitting on body through Imaged based -virtual try on. This project suggests a better solution using Artificial Intelligence (AI) and Augmented Reality (AR) for non-tech-savvy customers who aims at transferring a target clothing image onto himself. Creating realistic try-on images is still a significant obstacle, especially when dealing with extensive occlusions and complex human poses in the reference person. The algorithm determines whether to generate or preserve the image content based on the anticipated semantic layout. This approach results in highly realistic try-on images and captures intricate clothing details. This project idea is applying some steps below to get exact results. Beginning, based on the initial pose of the given person our model adjusts the target clothing form to compatible with the given pose. After this next task is to come up with the body segmentation model of the person wearing the required clothing, to better understand the body parts and clothing regions. Finally, the body segmentation map is fused together with warped clothing and a given person image for fine-scale image synthesis. The body segmentation map prediction using CNN, helps to guide image synthesis where body part and clothing intersects and it’s useful to preserving clothing and body part details.
:Augmented Reality (AR), Artificial Reality (AI), Virtual Try-On (VTON), Photo-Realistic Image