Face Recognition Based on Deep Learning Convolutional Neural Network in Cloud Internet of Things Environment
Year: 2025
Type: article
Abstract: Through the domain of remote medical engineering, the implementation of online facial recognition and intelligent routing has facilitated significant transformations across numerous sectors. In order to harness these functionalities, it is imperative that algorithms are both optimized and rendered efficient. Facial recognition within the contexts of the Internet of Things can be leveraged in domains such as security, healthcare, and surveillance. Nonetheless, in uncontrolled environments, the accuracy of facial recognition typically does not reach its full potential due to factors such as lighting conditions, angles of view, and the extent to which the face is occluded. For applications in healthcare and real-time systems, it is critical to achieve facial recognition that operates instantaneously and with high precision. The present study proposes the adoption of a deep learning-based algorithm structured in a tree configuration to sustain the actual accuracy of facial recognition in uncontrolled scenarios. Within such environments, this algorithm enhances recognition accuracy by partitioning the input volume into numerous smaller segments. The sequence in which the volume is divided is determined by the outputs of the deep learning algorithm, leading to a reduction in error rates in subsequent feedback iterations. This methodology supports deep learning processes and enhances the speed of algorithmic iteration cycles. The findings suggest that the proposed algorithm, which holds utility for real-time applications in security, healthcare, and surveillance, yields superior facial recognition accuracy in uncontrolled environments when compared to alternative methodologies.
Source: SN Computer Science
Institution: Islamic Azad University Sanandaj Branch
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,190
OpenAlex ID: W7117306829