A New Retexturing Method for Virtual Fitting Room Using Kinect 2 Camera

Andres Traumann, Gholamreza Anbarjafari, Sergio Escalera; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2015, pp. 75-79

This research work proposes a new method for garment retexturing using a single static image along with depth information obtained using the Microsoft Kinect 2 camera. First the garment is segmented out from the image and texture domain coordinates are computed for each pixel of the shirt using 3D information. After that shading is applied on the new colours from the texture image by applying linear stretching of the luminance of the segmented garment. The proposed method is colour and pattern invariant and results in to visually realistic retexturing. The proposed method has been tested on various images and it is shown that it generally performs better and produces more realistic results compared to the state-of-the-art methods. The proposed method can be an application for the virtual fitting room.