Logo Recognition using SURF Features and kNN Search Tree [ ]


In the contest of brand value and identity, the logo represents the company and gives strong impact on its reputation. So it is quite challenging to recognize a logo to maintain its standard level while designing. In this work, only SURF features are considered for logo recognition on gray scale image data. kNN search tree algorithm is used to get the nearest neighbors distance of SURF features between two different images. The percentage of matching features is calculated to recognize logo successfully. This percentage matching measurement of different data sets leads to recognize the logo of any organization undoubtedly when reference image dataset contains that logo.