Face Detection and Emotion Recognition Using Convolutional Neural Network

Face detection and emotion recognition have gained significant attention in the context of human-computer interaction and intelligent systems. This paper presents a Convolutional Neural Network (CNN)-based approach to automatic facial emotion recognition using a publicly available dataset of facial expressions. Images were preprocessed to a uniform resolution of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$48 \times 48$</tex> pixels and augmented to enhance model robustness. The CNN architecture integrates convolutional, pooling, batch normalization, and fully connected layers with dropout to enhance generalization. The model achieved a training accuracy of 87 % and a validation accuracy of approximately 70 %, highlighting both the promise and limitations of CNNs for this task. Results indicate effective capture of discriminative facial features but also highlight challenges in handling subtle and overlapping emotions such as fear and surprise. The findings provide a foundation for future work aimed at improving accuracy through deeper networks, attention mechanisms, and multimodal data integration. Our model achieved <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\sim 70 \%$</tex> validation accuracy, which is <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$+\mathrm{X} \%$</tex> higher than a simple CNN baseline without batch normalization and dropout.