Emotion Detection in Real-Time Using Computer Vision Techniques

In recent years, the field of human-computer interaction has made significant strides with the application of emotional intelligence in systems. Understanding and appropriately responding to human emotion can greatly enhance user experience in education, healthcare, marketing, and entertainment sectors. The aim of this project is to do real-time facial emotion detection using Convolutional Neural Networks (CNN) with live image processing through OpenCV. The model is trained on FER dataset which provides numerous facial images with corresponding expressions representing different emotions. Using Flask along with HTML and CSS, a GUI was created that allows for real-time emotion prediction from webcam video feed. The model demonstrates an acceptable performance of accurately detecting 68-70% of the time, which is deemed effective for the recognition of common emotions such as happy, sad, fear, and neutral. This paper presents the implemented workflow, the challenges encountered, analysis of results, and the additional steps needed to enhance system reliability and interactivity. Real-time emotion recognition enables emotionally intelligent systems that support mental health monitoring and enhance personalized learning experiences.