Scholarlink

<p>Early recognition and classification of brain tumors are significant areas in medical imaging that aids for providing mainly appropriate surgery to store patients' lives. Conventional machine learning techniques face significant challenges in achieving accurate tumor detection within minimal time consumption. Therefore, Neutrosophic Morphological Segmented Gaussian Regressive Deep Convolutional Network (NMSGRDCN) method is developed for efficient brain tumor detection with minimal time consumption. Medical brain MRI images serve as input and are transmitted to the input layer. An input medical image is sent to convolutional layer, wherever image denoising process is carried out using the Levenberg–Marquardt minimum variance smoothing filter to remove noisy artifacts. Subsequently, the preprocessed medical image is transmit to max pooling layer, wherever Neutrosophic Hamann Indexive Morphological Pixel Image Segmentation Process is carried out to divide the preprocessed image to segments and extract ROI. Finally, segmented image is forwarded to fully connected layer. Gaussian process regression is employed to analyze the extracted features, and binary step activation is used to categorize images as normal or tumor. The classified outcomes are attained at output layer. Experimental assessment is conducted using medical images, considering various metrics such as PSNR, disease diagnosis accuracy, precision, recall, F1 score, diagnosis time. Performance result shows that the NMSGRDCN technique improves data accuracy of brain tumor disease detection with reduced processing time compared to conventional methods.</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

Neutrosophic Morphological Segmented Gaussian Regressive Deep Convolutional Network for MRI Images Brain Tumor Classification

  • 期刊论文

  • 发表日期：

    2025年1月25日

查看全文
  • G Mohanapriya,
  • D Aarthi,
  • S Muthukumar,
  • M M Shanmugapriya,
  • S Santhosh Kumar

摘要

Early recognition and classification of brain tumors are significant areas in medical imaging that aids for providing mainly appropriate surgery to store patients' lives. Conventional machine learning techniques face significant challenges in achieving accurate tumor detection within minimal time consumption. Therefore, Neutrosophic Morphological Segmented Gaussian Regressive Deep Convolutional Network (NMSGRDCN) method is developed for efficient brain tumor detection with minimal time consumption. Medical brain MRI images serve as input and are transmitted to the input layer. An input medical image is sent to convolutional layer, wherever image denoising process is carried out using the Levenberg–Marquardt minimum variance smoothing filter to remove noisy artifacts. Subsequently, the preprocessed medical image is transmit to max pooling layer, wherever Neutrosophic Hamann Indexive Morphological Pixel Image Segmentation Process is carried out to divide the preprocessed image to segments and extract ROI. Finally, segmented image is forwarded to fully connected layer. Gaussian process regression is employed to analyze the extracted features, and binary step activation is used to categorize images as normal or tumor. The classified outcomes are attained at output layer. Experimental assessment is conducted using medical images, considering various metrics such as PSNR, disease diagnosis accuracy, precision, recall, F1 score, diagnosis time. Performance result shows that the NMSGRDCN technique improves data accuracy of brain tumor disease detection with reduced processing time compared to conventional methods.