Enhancing Social Media Face Emotion Recognition Using a Fuzzy ELM Approach

Our computational tools for searching, exploring, and sharing personal photographs are falling behind the rapid pace at which these images are being captured digitally. The use of automatic face recognition to categorize images according to the people in them is one potential solution. Accurate recognition on the Internet scale, however, presents the seemingly insurmountable challenge of picking out specific people from a pool of hundreds of millions. If large-scale face recognition is to be successful, this article contends that social network context is crucial. We may take advantage of the infrastructure and resources of online social networks to increase facial recognition rates on shared images, since many people post personal photos on the web through these sites. The four primary components of the suggested improved recognition system are the following: data collecting, data preprocessing, feature selection, Model training, and emotion recognition. In order to do multiclass classification, a number of NN techniques are employed. When optimising hyperparameters, the Group Wolf Optimisation (GGWO) algorithm is employed, and when selecting features, the Grey Wolf Optimisation (GWO) algorithm is used. When compared to existing state-of-the-art studies, the suggested model outperforms them with an accuracy of 98.56% for emotion recognition, and when utilising the FuzzyELM algorithm. The suggested FuzzyELM model works quite well, with an accuracy of 98.56%, a precision of 93.35%, a recall of 90.7%, and an F1-score of 91.93%. These numbers show that the model can accurately identify and categorise emotions based on facial expressions on social media. Despite its flaws, social media can be useful if we can decipher people's emotions from their postings, tweets, etc. As an example, it can deduce a person's intentions just before they take their own life. Recent research shows that the vast majority of suicide perpetrators post notes threatening suicide on social media; these messages should be treated with the seriousness they deserve.