Scholarlink

<p>In the context of evolution, human behavior can be termed as the way the individual interacts with the environment. Human behavior analysis is essential in today’s world to understand the way in which individuals respond to various scenarios and is mainly utilized for safety purposes, providing special education catering to the needs of people suffering from diseases, like Parkinson’s, Alzheimer and so on. This paper presents an automated human behavior recognition system, which recognizes human behavior based on fingerprint images. Here, the fingerprint pattern is first identified from the fingerprint images using the LeNet based on several texture features, minutiae features, and the enhanced input image. Further, the identified fingerprint patterns are compared with the data in the behavioral dictionary to recognize human behavior. Here, the parameters of the LeNet are adjusted using the Hybrid Gannet Bald Leader Optimization (HGBLO) algorithm. Furthermore, the superiority of the HGBLO_LeNet is examined in view of different evaluation metrics, like accuracy, False Positive Rate (FPR), True Positive Rate (TPR), Negative Predictive Value (NPV), Positive Predictive Value (PPV), False Negative Rate (FNR), and True Negative Rate (TNR), and the values attained are 0.935, 0.068, 0.913, 0.903, 0.912, 0.087, and 0.927, correspondingly.</p>

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

Hybrid Gannet Bald Leader Optimization Enabled LeNet for Fingerprint Pattern Classification for Human Behavior Analysis

  • 期刊论文

  • 发表日期：

    2026年2月16日

查看全文
  • Mokal Atul Bhimrao,
  • Brijendra Gupta

摘要

In the context of evolution, human behavior can be termed as the way the individual interacts with the environment. Human behavior analysis is essential in today’s world to understand the way in which individuals respond to various scenarios and is mainly utilized for safety purposes, providing special education catering to the needs of people suffering from diseases, like Parkinson’s, Alzheimer and so on. This paper presents an automated human behavior recognition system, which recognizes human behavior based on fingerprint images. Here, the fingerprint pattern is first identified from the fingerprint images using the LeNet based on several texture features, minutiae features, and the enhanced input image. Further, the identified fingerprint patterns are compared with the data in the behavioral dictionary to recognize human behavior. Here, the parameters of the LeNet are adjusted using the Hybrid Gannet Bald Leader Optimization (HGBLO) algorithm. Furthermore, the superiority of the HGBLO_LeNet is examined in view of different evaluation metrics, like accuracy, False Positive Rate (FPR), True Positive Rate (TPR), Negative Predictive Value (NPV), Positive Predictive Value (PPV), False Negative Rate (FNR), and True Negative Rate (TNR), and the values attained are 0.935, 0.068, 0.913, 0.903, 0.912, 0.087, and 0.927, correspondingly.