Abstract: Filter banks such as the Gabor Filter (GF) are widely used to describe objects. The main disadvantage of the Gabor filter is that it constructs redundant and incompact filters that may decrease system recognition performance. The purpose of the current study primarily is to enhance the categorization problem through generalizing the GF method (GGF). The unsupervised machine learning algorithm, denoted by the k-means clustering algorithm is proposed to implement generalization on a GF set. To assess the performance of the proposed method, the standard GF is used as a benchmark. Furthermore, the first 20 classes and the overall classes from the dataset Caltech 101 have been utalized in the performance demonstration of the newly suggested method. Based on a single classifier and combination feature (Naive approach), the proposed GGF outperforms and shows higher potential results than the standard GF in describing objects.
Hayder Ayad, Loay Edwar George, Mamoun Jassim Mohammed, Raad Ahmed Hadi and Siti Norul Huda Sheikh Abdullah, 2019. An Efficient Approach for Visual Object Categorization based on Enhanced Generalized Gabor Filter and SVM Classifier. Journal of Engineering and Applied Sciences, 14: 5753-5761.