In this project, we explore the field of emotion recognition using deep learning techniques. We conducted an analysis of a selected dataset consisting of 748 images. Our focus was on evaluating the accuracy and efficiency of algorithms, namely MobileNet, DenseNet, and VGG in recognizing emotions. The DenseNet algorithm achieved an accuracy of 80%, while MobileNet V2 performed lower at 79.33%. The VGG 16 and VGG 19 models demonstrated accuracies of 77.67% highlighting the varying performance of models when it comes to recognizing emotions. Notably, MobileNet emerged as the leading performer, boasting an impressive accuracy rate of 81.33%. Developed a user-friendly interface where users can simply upload an image. The predicted emotion from the image will be displayed as the label, whether it is happiness, sadness, disgust, angry, or surprise. Furthermore, our research highlights the vital role of context-aware datasets, showcasing the nuanced understanding achieved through the integration of custom images.