Advances in facial expression recognition technologies for emotion analysis

With the rapid evolution of technology, automated facial expression recognition (AFER) is one of the emerging research frontiers, gaining increasing attention from computer vision scientists. Facial emotion recognition (FER) enables a wide range of human–machine interaction applications, including mental health monitoring, behavioral analysis, and virtual learning environments. This paper comprehensively reviews recent advancements in AFER, examining both traditional and deep learning (DL) based approaches. The traditional section outlines methods involving feature extraction techniques such as LBP and Gabor filters, along with classifiers like Support Vector Machine (SVM), Principal Component Analysis (PCA), and k-nearest Neighbors (k-NN). In contrast, the DL section explores models such as Convolutional Neural Networks (CNNs) and hybrid networks that offer enhanced accuracy and adaptability. Various public FER datasets, including JAFFE, CK + , and FER2013, are analyzed to demonstrate the scope of existing benchmarks. Furthermore, this study identifies key gaps in the current FER literature, such as limited robustness in real-world conditions, challenges in recognizing emotions, and the lack of standardized evaluation protocols across datasets. We also highlight recent improvements in accuracy, feature extraction, and real-time applicability. This review aims to assist both emerging and established researchers by offering insights into current trends, performance metrics, and unresolved challenges in FER, ultimately contributing to the development of more robust and adaptive emotion recognition systems.