Context-Dependent Logo Matching and Recognition

Abstract

This project describes about the design of a novel variational framework to match and recognize multiple instances of multiple reference logos in image archives. Reference logos and test images are seen as constellations of local features (interest points, regions, etc.) and matched by minimizing an energy function mixing: 1) a fidelity term that measures the quality of feature matching, 2) a neighbourhood criterion that captures feature co-occurrence/geometry, and 3) a regularization term that controls the smoothness of the matching solution. This project also introduce a detection/recognition procedure and study its theoretical consistency. Finally, it show the validity of this method through extensive experiments on the challenging MICC-Logos dataset. Our method overtakes, by 20%, baseline as well as state-of-the-art matching/recognition procedures.

Existing System

Logos are graphic productions that either recall some real world objects, or emphasize a name, or simply display some abstract signs that have strong perceptual appeal. Color may have some relevance to assess the logo identity. But the distinctiveness of logos is more often given by a few details carefully studied by graphic designers, semiologists and experts of social communication. The graphic layout is equally important to attract the attention of the customer and convey the message appropriately and permanently. Different logos may have similar layout with slightly different spatial disposition of the graphic elements, localized differences in the orientation, size and shape, or in the case of malicious tampering differ by the presence/absence of one or few traits. In most of the cases they are subjected to perspective transformations and deformations, often corrupted by noise or lighting effects, or partially occluded. Such images and logos thereafter have often relatively low resolution and quality. Regions that include logos might be small and contain little information. Logo detection and recognition in these scenarios has become important for a number of applications. Among them, several examples have been reported in the literature, such as the automatic identification of products on the web to improve commercial search-engines, the verification of the visibility of advertising logos in sports events, the detection of near-duplicate logos and unauthorized uses. Special applications of social utility have also been reported such as the recognition of groceries in stores for assisting the blind.

Proposed System

In this project, we present a novel solution for logo detection and recognition which is based on the definition of a “Context Dependent Similarity” (CDS) kernel that directly incorporates the spatial context of local features [40], [41]. The proposed method is model-free, i.e. it is not restricted to any a priori alignment model. Context is considered with respect to each single SIFT key point and its definition recalls shape context with some important differences: given a set of SIFT interest Points X, the context of x∈X is defined as the set of points spatially close to x with particular geometrical constraints. Formally, the CDS function is defined as the fixed-point of three terms: (i) an energy function which balances a fidelity term; (ii) a context criterion; (iii) an entropy term. The fidelity term is inversely proportional to the expectation of the Euclidean distance between the most likely aligned interest points. The context criterion measures the spatial coherence of the alignments: given a pair of interest points(fp, fq)respectively in the query and target image with a high alignment score, the context criterion is proportional to the alignment scores of all the pairs close to (fp, fq) but with a given spatial configuration. The “entropy” term acts as a smoothing factor, assuming that with no a priori knowledge, the joint probability distribution of alignment scores is flat. It acts as a regularizer that controls the entropy of the conditional probability of matching, hence the uncertainty and decision thresholds so helping to find a direct analytic solution.

Architecture

ARCHITECTURE DIAGRAM