Lightweight Method for Two-Dimensional Object Registration for Augmented Reality

The purpose concerning this study is to develop an household guidance order based on augmented truth (AR) that superimposes annotations on an indoor environment (for instance, wall surfaces) and everything on devices with restricted computing power in the way that smartphones. Such a method demands a simple algorithm for the object enrollment. For the purpose, this paper proposes a two-dimensional (2D) object enrollment method established perspective distortion discipline using vanishing points in the scene and a fixed vanishing point discovery method. Current vanishing point discovery methods have the question of slight drift of the vanishing point in the videos despite claiming a fixed camera position. The projected method has resolved the problem by employing a burden particle filtering arrangement. Recent approaches often engage deep neural networks (DNNs) for the registration and the constant vanishing point detection, nevertheless DNNs require rich possessions for computing such as Graphics Processing Unit (GPU). Since the projected method requires merely two vanishing points for the registration, the procedure can be said a inconsequential method. This paper shows that the proposed burden particle percolating method suppresses the where earth meets sky drift better than the conventional moving average arrangement, and that it allows registration of objects based on the plan.

Author(s) Details:

Kazumoto Tanaka,
Kindai University, Japan.

Please see the link here: https://stm.bookpi.org/COSTR-V10/article/view/8905

Keywords: Augmented reality, object registration, vanishing point, particle filter