Scholarlink

The automatic tessellated fundus detection approach we provide in this paper makes use of texture and color information. The color fundus image is described using a combination of color moments, Local Binary Patterns (LBP), and Histograms of Oriented Gradients (HOG). The tessellated fundus is identified using an SVM classifier that has been trained after feature extraction. This paper employs and contrasts both linear and RBF kernels. To assess the suggested strategy, an 836-fundus image collection is created. The mean accuracy for the linear SVM is 98%, with sensitivity and specificity values of 0.99 and 0.98, respectively. The average RBF kernel accuracy is 97%, with 0.99 sensitivity and 0.95 specificity. According to the detection results, color and texture attributes can accurately represent the fundus images as normal or tessellated.

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

A Precision Tessellated Fundus Detection: Leveraging Color and Texture Features with SVM Classification

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Kachi Anvesh,
  • Bharati M. Reshmi,
  • Saroj Kumar Rout,
  • Bijaya Kumar Sethi,
  • Satya Sobhan Panigrahi

摘要

The automatic tessellated fundus detection approach we provide in this paper makes use of texture and color information. The color fundus image is described using a combination of color moments, Local Binary Patterns (LBP), and Histograms of Oriented Gradients (HOG). The tessellated fundus is identified using an SVM classifier that has been trained after feature extraction. This paper employs and contrasts both linear and RBF kernels. To assess the suggested strategy, an 836-fundus image collection is created. The mean accuracy for the linear SVM is 98%, with sensitivity and specificity values of 0.99 and 0.98, respectively. The average RBF kernel accuracy is 97%, with 0.99 sensitivity and 0.95 specificity. According to the detection results, color and texture attributes can accurately represent the fundus images as normal or tessellated.