Scholarlink

<p>To diagnose diseases and formulate treatment plans, the discipline of biomedical image screening is essential. As medical imaging technology develops further, it becomes more important than ever to guarantee the accuracy of these examinations. There are several restrictions on improved reliability in biomedical image screening, mostly because of issues with picture quality, dataset variety, and interpretability. This paper proposes a novel method based on bat-optimized VGG16 (BO-VGG16) for enhanced reliability in biomedical image screening. We begin by collecting brain MRI image datasets and preprocessing them using noise elimination based on bilateral filter and enhancement of images using histogram equalization. Furthermore, we extract relevant features from the preprocessed data using Linear Discriminant Analysis (LDA). A bat optimization algorithm is used to accelerate the proposed technique. Using a Python software tool, the experiment regarding brain MRI images was carried out. The result for tumor and no tumor classification, Accuracy (97.47% and 98.44%), Precision (94.02%, and 98.9%), Recall (97.35%, and 98.04%), F1-Score (95.65%, and 96.77%), FDR (0.0257, and 0.0103), TNR (0.0252, and 0.0186), metrics are optimized for the experimental findings. Experiment results show that the proposed (BO-VGG16) methodology performs superior to conventional approaches.</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

Improved medical science image filtering with a novel neural network technique

  • 期刊论文

  • 发表日期：

    2025年1月8日

查看全文
  • Khushboo Sharma,
  • Heli Amit Shah,
  • Nabanita Choudhury,
  • B. R. Samapangi Rama Reddy

摘要

To diagnose diseases and formulate treatment plans, the discipline of biomedical image screening is essential. As medical imaging technology develops further, it becomes more important than ever to guarantee the accuracy of these examinations. There are several restrictions on improved reliability in biomedical image screening, mostly because of issues with picture quality, dataset variety, and interpretability. This paper proposes a novel method based on bat-optimized VGG16 (BO-VGG16) for enhanced reliability in biomedical image screening. We begin by collecting brain MRI image datasets and preprocessing them using noise elimination based on bilateral filter and enhancement of images using histogram equalization. Furthermore, we extract relevant features from the preprocessed data using Linear Discriminant Analysis (LDA). A bat optimization algorithm is used to accelerate the proposed technique. Using a Python software tool, the experiment regarding brain MRI images was carried out. The result for tumor and no tumor classification, Accuracy (97.47% and 98.44%), Precision (94.02%, and 98.9%), Recall (97.35%, and 98.04%), F1-Score (95.65%, and 96.77%), FDR (0.0257, and 0.0103), TNR (0.0252, and 0.0186), metrics are optimized for the experimental findings. Experiment results show that the proposed (BO-VGG16) methodology performs superior to conventional approaches.