In the last decade, a plethora of handcrafted-based scene image classification techniques have been proposed in the literature. Some of them are based on structural analysis, while some others exploit mutual information or perceptual characteristics. Nowadays, deep learning-based methods are widely used in several domains due to its ability to well fit the target directly from the scene image. In this paper, an impact on the performance of combining handcrafted features and Deep features for scene image classification is addressed. In order to utilize the benefits of handcrafted and deep features, we extract and combine these features for scene image classification, which helps to improve the accuracy of the classification. We extract the deep features of the images based on the classical Res-Net model and handcrafted local features are extracted using Binary Robust Invariant Scalable Key points (BRISK) descriptor. By combining the two types of image features, we form a new type of image features, called hybrid features. Then, we employ Support Vector Machine (SVM) training model for scene image classification. We carried out the experiments on two public datasets such as CIFAR-10 and Corel-1000. The experimental results have shown that the proposed hybrid method has exhibited remarkable performance with high classification accuracy.

#