Automatic Facial Expression Recognition Employing CNN-Inception and the ResNet 50

The ability of computers to accurately determine the human state of mind can be improved with the help of the recognition of human facial expressions that reflect emotions. This will make it possible for workplace interactions to be more tailored to the individual. We will be able to recognize the emotions by paying attention to the placement of the brows and eyes, as well as the position of the mouth and the various changes that occur in the facial features. Using a deep learning Convolutional Neural Network (CNN)-INCEPTION and RESNET50 network model, we examine the recognition of human face expressions. The system utilizes a labeled data set that has around 32,298 photos with a variety of facial expressions. This data set is used for both training and testing purposes. A CNN model is trained using the dataset’s grayscale photos to categorize facial expressions into one of five categories: happy, sad, neutral, fearful, or furious. The use of batch normalization and dropout helps to increase the accuracy of the model while also preventing it from becoming overfit.