Image Preprocessing Method Applied on Facial Emotion Recognition Problem
Mingyang Xu · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2024
Sometimes people are unable to accurately express their views on a certain product, which may limit in accurately describing emotions.Facial emotion recognition can help solve the problem of people being unable to accurately express their opinions on products, by analyzing facial expressions to obtain emotional feedback.The application value and importance of this technology lies in providing a nonverbal way to understand user emotions, helping manufacturers enhance user satisfaction.Vision Transformer (ViT) model is very powerful on the Computer Vision Problems.In this paper.I used the VIT model to analyze the facial expression dataset and attempted different preprocessing methods.Through the VIT model.I can associate each data sample with the corresponding expression to determine the category of the expression.Then, compared the results obtained using different preprocessing methods and ultimately determine which method performs best in terms of accuracy.After comparing the accuracy obtained through various preprocessing methods, I found that Ran-domCrop has the highest accuracy and is most suitable for facial expression