Oppositional Grass Hopper Optimization with Fuzzy Classifier for Face Recognition from Video Database

Recognizing faces from the video employs artificial intelligence-based computer technology. This includes applications such as law enforcement, biometrics, security, personal safety, etc., for tracking and enabling surveillance in real-time. However, detection and recognition of the face from the video are influenced by the change in variation of pose, brightness, occlusion, expression, and resolutions. While facial images are simple to detect, others may necessitate the use of specialized software. To address those challenges we propose an efficient face detection and recognition system with optimal features. Initially, the keyframes are extracted by the Key Frame Extraction method which utilizes Wavelet Information. Subsequently, the characteristics such as holo-entropy, appearance features, SURF feature, and multi-angle movement feature are extracted. The Oppositional Grass Hopper Optimization Algorithm is used to identify optimal features from this large feature set. The extracted features are classified using the fuzzy classifier that was designed. The proposed method's performance is validated using several benchmark video datasets and its efficiency is measured in terms of keyframe extraction time, sensitivity, accuracy, and specificity by comparing it to other state-of-the-art approaches. The proposed method outperforms previous state-of-the-art methods in terms of recognition accuracy.