A novel facial expression recognition (FER) system combining K-Nearest Neighbours (KNN) and Long Short-Term Memory (LSTM) algorithms is proposed to enhance accuracy and efficiency in recognizing facial expressions. The system involves two main steps: feature extraction and classification. For feature extraction, the Local Binary Patterns (LBP) algorithm is used to capture texture information from facial images. In the classification stage, KNN is employed to classify facial expressions based on the nearest neighbours, while LSTM is used to handle temporal information in sequences of facial images. The LSTM processes LBP features through a series of cells to estimate the final expression. The system was tested on the CK+ and Oulu-CASIA datasets, achieving high performance with an F1 score and precision better than existing methods. The system's advantages include handling non-linear relationships, temporal dependencies, and achieving high accuracy even with limited data. The methodology involves data collection, preprocessing, feature extraction, model selection, training, and deployment. The system has potential applications in human-computer interaction, healthcare, security, marketing, and education. The study also discusses related research, including other FER methods and their performance on various datasets. The results show that the proposed system outperforms existing methods in terms of accuracy and error rate. The system's effectiveness is demonstrated through experiments on multiple datasets, showing high sensitivity, specificity, precision, and F1 scores for different emotions. The conclusion highlights the potential of the proposed FER system using KNN and LSTM for accurate and reliable facial expression recognition.A novel facial expression recognition (FER) system combining K-Nearest Neighbours (KNN) and Long Short-Term Memory (LSTM) algorithms is proposed to enhance accuracy and efficiency in recognizing facial expressions. The system involves two main steps: feature extraction and classification. For feature extraction, the Local Binary Patterns (LBP) algorithm is used to capture texture information from facial images. In the classification stage, KNN is employed to classify facial expressions based on the nearest neighbours, while LSTM is used to handle temporal information in sequences of facial images. The LSTM processes LBP features through a series of cells to estimate the final expression. The system was tested on the CK+ and Oulu-CASIA datasets, achieving high performance with an F1 score and precision better than existing methods. The system's advantages include handling non-linear relationships, temporal dependencies, and achieving high accuracy even with limited data. The methodology involves data collection, preprocessing, feature extraction, model selection, training, and deployment. The system has potential applications in human-computer interaction, healthcare, security, marketing, and education. The study also discusses related research, including other FER methods and their performance on various datasets. The results show that the proposed system outperforms existing methods in terms of accuracy and error rate. The system's effectiveness is demonstrated through experiments on multiple datasets, showing high sensitivity, specificity, precision, and F1 scores for different emotions. The conclusion highlights the potential of the proposed FER system using KNN and LSTM for accurate and reliable facial expression recognition.