Detection of positions and recognition of brand logos visible on images captured using mobile devices

Abstract

Up till now there does not exist an easy, mobile mechanism that allows to easily capture, recognize and count defined, multiple objects that are visible in surroundings of the user. For this purpose, feature detectors (such as SIFT, SURF or BRISK) are utilized to create a database of products box images and extracted keypoints are stored. Existing algorithms based on keypoints analysis do not allow to identify multiple identical logos, due to the fact that a homography calculated on found keypoints can span two or more objects and the result then can be skewed. In this paper a solution to this problem will be shown, that by using a sliding window that joins multiple found keypoints into individual objects, it is possible to correctly detect multiple identical objects. In this paper preliminary results of a mobile framework that allows recognition and counting of visible products in surroundings of the user will be presented.