A biometric identification system that analyzes body movements can increase interactivemodeling, robotics, virtual-authenticity, and biometric-identity through automatic recognition of emotions. A computer network, that deduces human emotions from body language, significantly alters how people interact with machines. Identification of emotion-specific features in numerous descriptions of human body movements is challenging. In this study, Deep BiLSTM system created using CMO method successfully detected human feelings. Gaits, movies, or feelings that were tagged serve as initial input for suggested emotion recognition-based method. For training, BiLSTM classifier is fed the most relevant features that were extracted and concatenated. Proposed Chiropteran Mahi optimization identifies better results for identifying revealed emotions of the classifier by tuning hyper parameters effectively. Learned videos and testing database are used to determine validity of proposed technique and educate it. Suggested approach has increased accuracy by 12.951% (depending on training%) and 7.793% based on KFold value compared with current approaches