Automatic Classification of Basic Emotions Using Deep Learning Techniques

Authors : Özen Özer, Nadir Subaşı

Pages : 75-88

View : 79 | Download : 197

Publication Date : 2025-12-23

Article Type : Research Paper

Abstract :This study aims to develop an advanced artificial intelligence system capable of automatically classifying seven basic emotions (anger, disgust, fear, happiness, neutrality, sadness, and surprise) through facial expressions. Utilizing Long Short-Term Memory neural networks, the system is designed to capture temporal variations in emotional expressions with high accuracy, robustness, and scalability. During the model development process, dataset diversity was ensured, data augmentation techniques such as rotation, cropping, and brightness adjustments were applied, and transfer learning was incorporated to enhance learning efficiency. The study thoroughly examines the impact of data organization on model performance and analyzes how different data representation methods affect accuracy rates. Experimental results demonstrate that the Long Short-Term Memory based architecture effectively captures temporal dynamics in facial expressions, outperforming traditional methods in emotion recognition tasks. The system’s real-time processing capability makes it suitable for applications in healthcare, education, and security. Ethical considerations, including data privacy, informed consent, and bias mitigation, have been prioritized to ensure fair and responsible AI deployment. The findings highlight the significant potential of emotion recognition technology in human-computer interaction and emphasize the need for future research on multimodal emotion recognition, integration of diverse data sources, and the establishment of ethical guidelines to prevent misuse.
Keywords : duygu sınıflandırma, derin öğrenme, uzun kısa süreli bellek, yüz ifade analizi, yapay zeka, çeşitli veri kümesi

ORIGINAL ARTICLE URL