Logo Matching and Recognition Based on Context

Logo Matching and Recognition Based on Context

Tapan Kumar Das (VIT University, India)
Copyright: © 2018 |Pages: 13
DOI: 10.4018/978-1-5225-5775-3.ch009
OnDemand PDF Download:
No Current Special Offers


Logos are graphic productions that recall some real-world objects or emphasize a name, simply display some abstract signs that have strong perceptual appeal. Color may have some relevance to assess the logo identity. Different logos may have a similar layout with slightly different spatial disposition of the graphic elements, localized differences in the orientation, size and shape, or differ by the presence/absence of one or few traits. In this chapter, the author uses ensemble-based framework to choose the best combination of preprocessing methods and candidate extractors. The proposed system has reference logos and test logos which are verified depending on some features like regions, pre-processing, key points. These features are extracted by using gray scale image by scale-invariant feature transform (SIFT) and Affine-SIFT (ASIFT) descriptor method. Pre-processing phase employs four different filters. Key points extraction is carried by SIFT and ASIFT algorithm. Key points are matched to recognize fake logo.
Chapter Preview

Early work on logo detection and recognition was concerned with providing some automatic support to the logo registration process. The system must check whether other registered logos in archives of millions, exist that have similar appearance to the new coming logo image, in order to ensure that it is sufficiently distinctive and avoid confusion. Kato’s system was among the earliest ones. It mapped a normalized logo image to a 64 pixel grid, and calculated a global feature vector from the frequency distributions of Edge pixels.

A method was proposed by Lowe (2004) for extracting invariant features from images that can be used to perform reliable matching between different views of an object. The recognition was done by matching individual features to a database of features from known objects using a fast nearest-neighbor algorithm (Das & Chowdhury, 2017) which was followed by a hough transform to identify clusters belonging to a single object. In another embodiment, Ballan et al proposed trademark detection and recognition system while advertising trademark in a sports videos (Ballan, Bertini & Jain, 2008). It is a semi-automatic system for detecting and retrieving trade-mark appearances in a sports videos. A human annotator supervises the results of the automatic annotation through an interface that shows the time and the position of the detected trademarks.

Complete Chapter List

Search this Book: