Facial expression recognition (FER) is a common investigation phenomenon, resulting in several computational vision tasks like super-resolution reconstruction, image generation, image translation, and video generation. FER has led to many challenges, which have resulted in numerous technological advances in AI and computer vision. The major steps of FER include preprocessing, feature extraction, and classification. Several face detection techniques have been proposed, including the eigenspace technique, Viola-Jones, and adaptive skin color algorithm among others, and the methods have been designed centered on the Haar classifier, contour points, and Adaboost. The key methods used in feature extraction include principal component analysis, Gabor feature, local binary patterns (LBPs), and active appearance models. Various classification algorithms are deployed in this phase, including directed Line segment Hausdorff Distance (dLHD), K-nearest neighbors (KNN), Support Vector Machine, Hidden Markov Model (HMM), Hidden Conditional Random Fields (HCRF), Online Sequential Extreme Learning Machine (OSELM), Learning Vector Quantization (LVQ), ID3 Decision Tree (DT), Multilayer Feed Forward Neural Network (MFFNN), Bayesian Neural Network, Convolution Neural Network (CNN), Deep Neural Network (DNN), and Deep Belief Network (DBN). Notwithstanding the recent development, FER is still a challenging task influenced by several factors like pose variation, multifaceted background, illumination, and pose variation.