Deep Learning for Logo Classification

Abstract

Logos are the brand ambassadors of any company or the organization which serves as the identity of it in many of the ways. Logos plays a distinct and significant role providing information about the documents of the particular organizations. Logos can be classified based on the size, texture, color, information, shapes and many other factors. In this work here, we mainly focus on classification of different types of logos in two stages of classification systems using Deep learning and KNN with Gabour and SIFT features respectively. The advantage of the proposed method is that the two stage classification yields better results and the accuracy with the classification of testing images is high.