Tri Bird Technique for Effective Face Recognition Using Deep Convolutional Neural Network

Abstract

The massive amount of data produced from Video content, has speedily grown up with surveillance data. Human beings can easily handle this data by greatly outperforming the capacity of technical and non-technical resources. In the current era of computer vision great characteristics like precision, effectiveness and interoperability of smart surveillance systems to retrieve data. Security surveillance systems rapidly increase Face value acknowledgement over different subsequent critical methods. In proposed research, deep CNN and deep BiLSTM methods were used for effective face recognition from video input. Generally, the said techniques efficiently handle challenges face recognition by combining the recompenses of computer vision and tri bird optimization. The proposed method can be used in many computer vision applications that help recognize faces. The hybrid deep model is developed using Distributed Deep CNN classifier and distributed BiLSTM classifier, which is optimized using the new optimization that is designed newly for your work. The new optimization is designed based on the characters of vulture, sparrow and crow. Then the output of the classifier is fed through the model by analyzing the face query to attain the retrieved relevant face. The outcome can be measured by enactment over the proposed supposed model and is analyzed through considering metrics such as accuracy, exactness, remembrance and F-value calculator which are implemented in the MATLAB tool to reveal the proposed face recognition method. The proposed model achieves 91.30% accuracy, 95.07% precision, 95.078% recall and 95.078% f-measures by deep convolutional neural network.