Scholarlink

In the current, face recognition has become the most sought-after technology, and the one most integrated into the technologies used in various fields such as security, video surveillance and authentication systems. This growing demand for this technology is an incentive to improve algorithms and innovate reliable and robust facial recognition approaches for this application. The aim of this study is to introduce a new approach to facial recognition using the estimation landmark algorithm for face detection in videos or images. Next, a new architecture of a convolutional neural network based on a parameterized function of the activation function Relu, this new function avoids the problems of “dead” neurons and gradient disappearance. This new CNN model enables better extraction of facial features and reliable, accurate classification.

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

Enhanced Facial Recognition Using Parametrized ReLU Activation in Convolutional Neural Networks

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Hassan Afifi,
  • Abdallah Marhraoui Hsaini,
  • Mostafa Merras,
  • Aziz Bouazi,
  • Idriss Chana

摘要

In the current, face recognition has become the most sought-after technology, and the one most integrated into the technologies used in various fields such as security, video surveillance and authentication systems. This growing demand for this technology is an incentive to improve algorithms and innovate reliable and robust facial recognition approaches for this application. The aim of this study is to introduce a new approach to facial recognition using the estimation landmark algorithm for face detection in videos or images. Next, a new architecture of a convolutional neural network based on a parameterized function of the activation function Relu, this new function avoids the problems of “dead” neurons and gradient disappearance. This new CNN model enables better extraction of facial features and reliable, accurate classification.