Gabor-Based Novel Color Descriptors for Object and Scene Image Classification

This paper presents several novel Gabor-based color descriptors for object and scene image classification. Firstly, a new Gabor-HOG descriptor is proposed for image feature extraction. Secondly, the Gabor-LBP descriptor derived by concatenating the Local Binary Patterns (LBP) histograms of all the component images produced by applying Gabor filters is integrated with the Gabor-HOG using an optimal feature representation method to introduce a novel Gabor-LBP-HOG (GLH) image descriptor which performs well on different object and scene image categories. Finally, the Enhanced Fisher Model (EFM) is applied for discriminatory feature extraction and the nearest neighbor classification rule is used for image classification. The robustness of the proposed GLH feature vector is evaluated using three grand challenge datasets, namely the Caltech 256 dataset, the MIT Scene dataset and the UIUC Sports