Classification of Forged Logo Images

In this paper, the Binary and Hierarchical Multi-classification models for forged logo classification based on texture features are proposed in the binary classification model where the texture of the logo is extracted using two different models like Grey Level Co-Occurrence Matrix (GLCM), and Local Binary Patterns (LBP). The K Nearest Neighbor (KNN) classifier and Support Vector Machine (SVM) classifiers are used for classification. In hierarchy approach, a two-level hierarchy is used for the classification of the forged Logos across the four predefined classes. At the first level, the system is checked for genuine and forgery logos using binary classification. After applying binary classification in the first level if the input logo belongs to the forged class then in second-level multi-classification is used to classify the forged logo across the predefined classes. Same feature extraction and classification methods are used in second-level classification. Experimentations on a dataset containing 5025 color logo images of 5 classes are tested to demonstrate the proposed model performance.