3D Camouflaging Object using RGB-D Sensors

TL;DR

This paper introduces a 3D optical camouflage system utilizing RGB-D sensors to conceal objects behind displays by estimating shadow shapes on surrounding 3D objects, enhancing concealment accuracy.

Contribution

It presents a novel method combining RGB-D data and eye tracking to predict shadow regions for improved 3D object concealment.

Findings

Successfully estimates shadow shapes on 3D backgrounds

Enables concealment of objects behind displays in complex scenes

Integrates eye tracking with point cloud analysis for real-time application

Abstract

This paper proposes a new optical camouflage system that uses RGB-D cameras, for acquiring point cloud of background scene, and tracking observers eyes. This system enables a user to conceal an object located behind a display that surrounded by 3D objects. If we considered here the tracked point of observer s eyes is a light source, the system will work on estimating shadow shape of the display device that falls on the objects in background. The system uses the 3d observer s eyes and the locations of display corners to predict their shadow points which have nearest neighbors in the constructed point cloud of background scene.