Application for contextual images classification

Abstract

The goal of this master’s thesis is to design, develop, and implement a comprehensive system that can effectively classify images based on their context. To achieve this objective, we employed two multimodal learning approaches, which enable us to capture and analyze long-term dependencies and contextual information more effectively. To demonstrate the performance of the proposed methods, experiments were conducted on a custom dataset. The evaluation of the chosen method yielded a classification accuracy of 80% Key words: Artificial intelligence, image classification, deep learning, contextual image classification, multimodal learning