Scholarlink

The main way to recognize someone is through their facial features, which may even tell apart identical twins. In order to separate people, facial recognition and identification become essential. Modern uses for this technology include phone unlocking systems, criminal identification systems, and home security systems. This technology is seen as more secure because it depends on a facial image rather than outside elements like a card or key. Face detection and face identification are the two main processes in the recognition process. This article explores the idea of creating a deep learning-based face recognition system using Python’s OpenCV module. Deep learning is the optimal approach for facial recognition because of its outstanding accuracy. This research paper presents an in-depth exploration into real-time person identification employing face recognition technology. The objective of this study is to develop a robust system capable of accurately identifying individuals in various real-time scenarios. Proposed system uses own developed dataset and architecture integrates the Jetson Nano platform, known for its cost-effectiveness and computational power, along with a web camera for image acquisition. Each algorithm plays a distinctive role in the face recognition process. The Histogram of Oriented Gradients (HOG) method contributes by providing robust feature descriptors, while the Support Vector Machine (SVM) ensures efficient classification. Additionally, the deep learning models, ResNet and CNN, further enhance recognition accuracy by capturing intricate facial features and patterns. The hardware results using NVIDIA Jetson nano demonstrate the system’s capability to achieve real-time person identification with high accuracy rates, rendering it suitable for deployment in security, access control, and surveillance applications where rapid and reliable identification is critical. The System achieves the 97% recognition accuracy for Resnet-50 architecture for each entry of identified person daily log is recorded furthermore.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development of Real Time Person Identification Using SVM and Resnet-50 Algorithm

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Meghana Deshpande,
  • Rupali Kamathe,
  • Vandana Hanchate,
  • Sheetal Borde,
  • Kalyani Joshi,
  • Purva Marwade,
  • Pratiksha K. Kamble,
  • Priti D. More

摘要

The main way to recognize someone is through their facial features, which may even tell apart identical twins. In order to separate people, facial recognition and identification become essential. Modern uses for this technology include phone unlocking systems, criminal identification systems, and home security systems. This technology is seen as more secure because it depends on a facial image rather than outside elements like a card or key. Face detection and face identification are the two main processes in the recognition process. This article explores the idea of creating a deep learning-based face recognition system using Python’s OpenCV module. Deep learning is the optimal approach for facial recognition because of its outstanding accuracy. This research paper presents an in-depth exploration into real-time person identification employing face recognition technology. The objective of this study is to develop a robust system capable of accurately identifying individuals in various real-time scenarios. Proposed system uses own developed dataset and architecture integrates the Jetson Nano platform, known for its cost-effectiveness and computational power, along with a web camera for image acquisition. Each algorithm plays a distinctive role in the face recognition process. The Histogram of Oriented Gradients (HOG) method contributes by providing robust feature descriptors, while the Support Vector Machine (SVM) ensures efficient classification. Additionally, the deep learning models, ResNet and CNN, further enhance recognition accuracy by capturing intricate facial features and patterns. The hardware results using NVIDIA Jetson nano demonstrate the system’s capability to achieve real-time person identification with high accuracy rates, rendering it suitable for deployment in security, access control, and surveillance applications where rapid and reliable identification is critical. The System achieves the 97% recognition accuracy for Resnet-50 architecture for each entry of identified person daily log is recorded furthermore.