The Comparison of Two Image Matching Algorithms Based on Real-Time Image Acquisition

Abstract

Compared with traditional printing, mobile phone device is easier to carry and has richer appearance of data presentation. For traditional printing, combining traditional printing with mobile phone to express more information which contained in the prints has become a new development direction. Matching the image acquired by mobile phone camera from traditional prints, if the matching is successful, the mobile phone will display animation, music, or video which associated with the print content. Hence, image matching algorithm plays a vital role in this process. Due to the constant need for real-time computation of the image, and limited capabilities of mobile devices, the matching algorithm’s speed, accuracy, and efficiency should be the first priority. This paper combines the FAST feature point’s detection algorithm with the FREAK feature point’s description algorithm to accomplish the function of detecting feature points. Then, the two algorithms were compared in different state. Experimental results show that ORB has high matching accuracy with slower speed. The FREAK’s speed is faster, and the accuracy performance can satisfy the demand of image matching.