A Comprehensive Approach for Scenes Classification

Abstract

Abstract Scene classification is one of the most complex tasks in computer vision. The accuracy of scene classification is dependent on other subtasks such as object detection and object classification. Machine and transfer learning are widely employed in scene classification achieving optimal performance. Despite the promising performance of existing models in scene classification, there are still major issues. First, the training phase for the models necessitates a large amount of data, which is a difficult and time-consuming task. Furthermore, most models are reliant on data previously seen in the training set, resulting in ineffective models that can only identify samples that are similar to the training set. As a result, few-shot learning has been introduced. We propose a scene classification system that can operate on various sizes of data sets and tend toward scene classification generalizability. Our experiments include a novel machine learning model for scene classification applied to Intel scenes data set with a novel reconstruction of the data to serve binary and multi-classifications tasks. Moreover, we developed three novel architectures that use few-shot learning for scene classification achieving optimal accuracy compared to the existing models in the literature. The proposed system includes models for few-shot learning that achieved accuracies of 52.16, 35.86, and 37.26 for five-shots on the MiniSun, MiniPlaces, and MIT-Indoor 67 benchmark datasets, respectively, while the proposed machine learning model achieved accuracies of 93.55, 75.54 for training and validation data of Intel scenes data.