Academic Journals Database
Disseminating quality controlled scientific knowledge

Real Time Implementation of HIPs Descriptor on Android Mobile Platform in an Augmented Reality Application

ADD TO MY LIST
 
Author(s): Marek Jakab

Journal: Information Sciences and Technologies Bulletin of the ACM Slovakia
ISSN 1338-1237

Volume: 5;
Issue: 2;
Start page: 28;
Date: 2013;
VIEW PDF   PDF DOWNLOAD PDF   Download PDF Original page

Keywords: Augmented Reality | Mobile Device | Descriptor | HIPs | Histogram | Image Recognition | Keypoint

ABSTRACT
The purpose of our research is to develop an applicationof augmented reality on mobile device, which is educativeand entertaining for its users – children. A user is asked tocapture images using a camera built in the mobile deviceand the application will provide a supplementary infor-mation into the image in the real time. The augmentedinformation could be for example a text or graphics inthe same image. Typically, the images are taken fromchildren books. The key task of our application is theproblem of image recognition on mobile platform usinglocal descriptors. Our goal is to develop an applicationwhich is running in real time. Currently available descrip-tors included in OpenCV library are well designed, someof them are scale and rotation invariant, but most of themare time and memory consuming and hence not suitablefor mobile platform. Therefore we decided to develop afast binary descriptor based on the Histogram of IntensityPatcheS (HIPs) originally proposed by Simon Taylor et al.To train the descriptor, we need a set of images derivedfrom a reference picture taken under varying viewing con-ditions and varying geometry. Our descriptor is based ona histogram of intensity of the selected pixels around thekey-point. We use this descriptor in the combination withthe FAST key-point detector and a sophisticated methodof key-points selection is then used with the aim to reducethe computation time.
Affiliate Program      Why do you need a reservation system?