Towards Image-OLAP Integration: Leveraging Deep Learning Based Facial Emotion Classification for Advanced Image Data Warehousing

This paper addresses the challenge of integrating structured and unstructured image data in decision support systems, as traditional OLAP systems struggle with unstructured data. It introduces the concept of Image-OLAP, extending OLAP tools to process both image and structured data. The proposed model incorporates Convolutional Neural Networks (CNNs) for efficient image retrieval, classification, and analysis, with facial emotion recognition (FER) as a contextual dimension to enhance interpretability. This approach is particularly useful in fields like healthcare, customer service, and human resources. The Image-OLAP framework uses a star schema for multidimensional analysis of both data types, providing a comprehensive decision-making tool. Experimental results validate the model's effectiveness in emotion classification and data integration. This research lays the groundwork for developing Image-OLAP systems that enable more informed decision-making across industries, advancing data analytics and decision support systems.