This paper presents an enhanced real-time facial expression recognition (FER) system using deep learning. The study focuses on the application of convolutional neural networks (CNNs) for the extraction and classification of seven universal emotional states: happiness, sadness, anger, fear, surprise, disgust, and neutrality. The research addresses the challenges of FER in real-time, uncontrolled environments, where factors such as degraded image quality, occlusions, variable lighting, and head pose variations can affect performance. The proposed method employs deep learning techniques to process images in real-time, particularly those of lower resolution, to improve the accuracy of FER in real-world datasets, which are inherently more complex and collected under less controlled conditions compared to laboratory-collected data. The system uses a deep learning-based approach for emotion detection in photographs, which is rigorously evaluated. The proposed method is compared with manual techniques and other existing approaches to assess its effectiveness. This comparison forms the foundation for a subjective evaluation methodology, focusing on validation and end-user satisfaction. The findings conclusively demonstrate the method's proficiency in accurately recognizing emotions in both laboratory and real-world scenarios, thereby underscoring the potential of deep learning in the domain of facial emotion identification. The proposed model is based on MobileNet, a variant of CNNs developed by Google, which is efficient in processing and has a reduced parameter count compared to conventional convolutional networks. The model is designed for real-time processing and is capable of differentiating between seven emotional categories. The system's architecture includes data collection, preprocessing, emotion prediction, and performance evaluation. The model's performance is evaluated using predefined datasets, and the results show a high accuracy of 97.9% during training and 97.7% during validation. The model's effectiveness is demonstrated through extensive experiments, showing its ability to accurately recognize emotions in various scenarios. The study highlights the importance of acknowledging the inherent challenges and limitations in emotion detection tasks, particularly with images exhibiting a range of similar emotional expressions. Future research directions include exploring additional classes, enhancing accuracy, and implementing real-time facial emotion recognition using cameras.This paper presents an enhanced real-time facial expression recognition (FER) system using deep learning. The study focuses on the application of convolutional neural networks (CNNs) for the extraction and classification of seven universal emotional states: happiness, sadness, anger, fear, surprise, disgust, and neutrality. The research addresses the challenges of FER in real-time, uncontrolled environments, where factors such as degraded image quality, occlusions, variable lighting, and head pose variations can affect performance. The proposed method employs deep learning techniques to process images in real-time, particularly those of lower resolution, to improve the accuracy of FER in real-world datasets, which are inherently more complex and collected under less controlled conditions compared to laboratory-collected data. The system uses a deep learning-based approach for emotion detection in photographs, which is rigorously evaluated. The proposed method is compared with manual techniques and other existing approaches to assess its effectiveness. This comparison forms the foundation for a subjective evaluation methodology, focusing on validation and end-user satisfaction. The findings conclusively demonstrate the method's proficiency in accurately recognizing emotions in both laboratory and real-world scenarios, thereby underscoring the potential of deep learning in the domain of facial emotion identification. The proposed model is based on MobileNet, a variant of CNNs developed by Google, which is efficient in processing and has a reduced parameter count compared to conventional convolutional networks. The model is designed for real-time processing and is capable of differentiating between seven emotional categories. The system's architecture includes data collection, preprocessing, emotion prediction, and performance evaluation. The model's performance is evaluated using predefined datasets, and the results show a high accuracy of 97.9% during training and 97.7% during validation. The model's effectiveness is demonstrated through extensive experiments, showing its ability to accurately recognize emotions in various scenarios. The study highlights the importance of acknowledging the inherent challenges and limitations in emotion detection tasks, particularly with images exhibiting a range of similar emotional expressions. Future research directions include exploring additional classes, enhancing accuracy, and implementing real-time facial emotion recognition using cameras.