As is well recognised, emotions have a significant impact on how information is processed, attitudes are formed, and decisions are made in real-world situations. Although there have been recent publications regarding FER, or facial expression recognition, developing accurate and stable FER systems is still difficult because of the variety of human faces and variations in images. Every study and piece of research up to this point has suggested either a single network or an ensemble model. The accuracy of ensemble models is better, but they require a greater number of models, datasets, and modified datasets, which adds to the computational complexity. While the majority of research in this area focuses on increasing accuracy, this work applies the suggested model to a situation when a person's face incorporates a variety of emotions. In such circumstances, a single-label sentiment can be quite noisy. In response to this circumstance, we created and evaluated 15–20 models and techniques. In this paper, we propose a single standalone-based CNN model with its implementation on a real-time Intelligent System for Sentiment Recognition. This system performs tasks like face detection, sentiment classification, and providing a live list of probabilistic labels in Realtime from a webcam feed in one blended step, validating accuracy through transfer learning. The suggested model outperforms all standalone-based approaches, such as VGG16, VGG19, EfficientNetB7, and other other models, with an accuracy will be good.hen applied alone to the difficult and noisy FER2013 dataset, which is a crowd-sourced dataset.
We smart ai technologies provide best final year engineering projects with complete support.
Below are the support we provide :
1 complete working code implementation in clients machines
2 complete classes on project will be held
3 documentation support
4 Help in researching without extra cost/price increase.