Analysis of RGB Depth Sensors on Fashion Dataset for Virtual Trial Room Implementation

This paper presents a Virtual Trial Room software using Augmented Reality which allows the user to wear clothes virtually by superimposing 3d clothes over the user. These sensors are valued particularly for robotics or computer vision applications because of their low cost and their ability to measure distances at a high frame rate. In November 2010, the Kinect v1 (Microsoft) release encouraged the use of Red Green Blue (RGB)-D cameras, and in July 2014, a second version of the sensor was launched. Because high-frequency point nuclei can be obtained from an observed picture, users can imagine employing these sensors to fulfill 3D acquisition requirements. However, certain issues such as the adequacy and accuracy of RGB-D cameras in close-range 3D modeling have to be addressed owing to the technology involved. The quality of the data obtained therefore constitutes an important dimension. In this study, the usage of the current sensor Kinect v2 is explored in the three-dimensional reconstruction of tiny objects. The advantages and problems of Kinect v2 are addressed in the first section and then photogrammetry versions are presented after an accurate evaluation of the generated models. 

Keywords: Augmented reality, Depth Sensor, Fashion Recommendation, IOT, Virtual reality, Virtual trial room.