Scholarlink

<p>Biometric identification through ear image analysis and recognition has emerged as a promising biometric system due ear uniqueness and stability in varying environmental conditions. In this study, a novel approach for ear recognition using a fusion of two powerful feature extraction techniques such as Histogram of Oriented Gradients (HOG) and an improved Local Optimal Oriented Pattern (LOOP) based on Spatial Pyramid Decomposition (SPD) is proposed. An efficient scheme based on Discriminant Correlation Analysis (DCA) is adopted as a feature level fusion and dimensionality reduction technique where the performance of the recognition system is evaluated using a K-Nearest Neighbors (K-NN) classifier. Extensive experiments on six well known benchmarks ear datasets are conducted to assess the effectiveness of the proposed approach. Experimental results clearly indicate the superiority of the proposed method in terms of performance and complexity in comparison with the state-of-the-art ear recognition techniques.</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

An improved ear recognition system based on efficient feature extraction and fusion techniques

  • 期刊论文

  • 发表日期：

    2025年9月22日

查看全文
  • Toufik Lebed,
  • Abdelhak Boukharouba

摘要

Biometric identification through ear image analysis and recognition has emerged as a promising biometric system due ear uniqueness and stability in varying environmental conditions. In this study, a novel approach for ear recognition using a fusion of two powerful feature extraction techniques such as Histogram of Oriented Gradients (HOG) and an improved Local Optimal Oriented Pattern (LOOP) based on Spatial Pyramid Decomposition (SPD) is proposed. An efficient scheme based on Discriminant Correlation Analysis (DCA) is adopted as a feature level fusion and dimensionality reduction technique where the performance of the recognition system is evaluated using a K-Nearest Neighbors (K-NN) classifier. Extensive experiments on six well known benchmarks ear datasets are conducted to assess the effectiveness of the proposed approach. Experimental results clearly indicate the superiority of the proposed method in terms of performance and complexity in comparison with the state-of-the-art ear recognition techniques.