A Deep Learning-Based Intelligent Face Recognition Method in the Internet of Home Things for Security Applications

Abstract

At the beginning of the 21st century, human life has become increasingly dependent on security. At this point, the cost is the essential consideration. This approach can reduce monitoring costs. This research proposes a real-time recognition system to deal with photos as soon as possible. We hope that by identifying people, we can keep our homes and offices safe. One of the primary goals of this research was to develop a method of intelligent face recognition for smart homes based on deep learning. To demonstrate the usefulness of our study, we also examine and contrast this model with other methodologies deemed contemporary. This research proposes a tree-based deep model for cloud-based face recognition. The proposed deep model is less computationally demanding without affecting accuracy. Trees for each volume are built in the model's input volume, split into many ones. The number of branches and their height characterizes trees. Residual functions are used to represent each branch, and they are built from a convolutional layer and two non-linear functions. In various openly accessible databases, the proposed model is put to the test. It also compares to the best deep facial recognition models in the industry today.