Scholarlink

<p>This paper presents a method for offline handwritten Gurmukhi character recognition that leverages the fusion of classifiers. The proposed system for recognizing isolated Gurmukhi handwritten characters combines various features through fusion. Principal Component Analysis (PCA) is used to extract efficient features for the classification process. For classifier fusion, both k-NN and SVM are employed. A key aspect of this study is the evaluation of the proposed methodology using a benchmark dataset of Gurmukhi script, which includes 5600 handwritten characters from one hundred writers. Experiments were conducted with different partitioning strategies and k-fold cross-validation techniques. The system achieved a recognition accuracy of 92.3% when 20% of the data was used as the test set.</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

Improving Character Recognition Using Classifier Fusion and Feature Selection

  • 期刊论文

  • 发表日期：

    2026年4月17日

查看全文
  • Kanta Prasad Sharma,
  • Ali Khatibi,
  • Krunal Vaghela,
  • Jayaprakash B.,
  • Pooja Rani,
  • Nilesh Bhosle,
  • M. Janaki Ramudu,
  • Devendra Singh

摘要

This paper presents a method for offline handwritten Gurmukhi character recognition that leverages the fusion of classifiers. The proposed system for recognizing isolated Gurmukhi handwritten characters combines various features through fusion. Principal Component Analysis (PCA) is used to extract efficient features for the classification process. For classifier fusion, both k-NN and SVM are employed. A key aspect of this study is the evaluation of the proposed methodology using a benchmark dataset of Gurmukhi script, which includes 5600 handwritten characters from one hundred writers. Experiments were conducted with different partitioning strategies and k-fold cross-validation techniques. The system achieved a recognition accuracy of 92.3% when 20% of the data was used as the test set.