Abstract: Logo matching and recognition is mainly done in order to reveal duplicate logos which have a slight difference from the original logos in order to deceive customer. In this paper an effective logo matching and recognition method called Context-Dependent Similarity method is performed. Firstly dataset of logo images are created for the reference images, then another test image which have logos in it is taken as the input and start matching with the dataset logo images to check whether the input image is fake or genuine logo based on the dataset image. For this features of logos images for both the reference and test images are extracted and based on their extracted features matching is performed using Context-Dependent Similarity method.

Keywords: Key-points extraction, Computation of context, Logo matching and recognition.