Scholarlink

<p>Prostate cancer is one of the most commonly occurring cancers in males and management and treatment depends on an accurate and timely diagnosis. This paper introduces a novel computer aided diagnosis (CAD) pipeline that utilizes optimized deep segmentation, advanced texture analysis, and hybrid classification to detect prostate cancer from MRI. The CAD pipeline begins with preprocessing which includes, but is not limited to, intensity normalization, Gaussian noise filtering, image rescaling, and advanced data augmentation for generalization. An optimized Unet (O-Unet) architecture is designed for high performance prostate and tumor segmentation and subsequently, an Improved Gray-Level Co-occurrence Matrix (IGLCM) is developed for discriminative feature extraction from the segmented partitions. In the second stage of our framework, we explore three possibilities for classification: (i) Artificial Neural Network (ANN) using statistical features, (ii) Convolutional Neural Network (CNN) utilizing spatial features, and (iii) a hybrid ANN–CNN model that can take advantage of the strengths of both. Experimental results demonstrate that the task would be completed with increased diagnostic accuracy and robustness to MRI variances, and my clinical decision support would exceed existing CAD. The proposed CAD system performs accurate prostate and lesion segmentation followed by robust classification of prostate cancer using a hybrid ANN–CNN model.</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

Prostate cancer detection from MRI using optimized UNET and hybrid ANN–CNN classification for enhanced diagnosis

  • 期刊论文

  • 发表日期：

    2026年6月18日

查看全文
  • V. P. Gladis Pushparathi,
  • D. Jasmine David,
  • R. Ramesh,
  • S. Padmapriya

摘要

Prostate cancer is one of the most commonly occurring cancers in males and management and treatment depends on an accurate and timely diagnosis. This paper introduces a novel computer aided diagnosis (CAD) pipeline that utilizes optimized deep segmentation, advanced texture analysis, and hybrid classification to detect prostate cancer from MRI. The CAD pipeline begins with preprocessing which includes, but is not limited to, intensity normalization, Gaussian noise filtering, image rescaling, and advanced data augmentation for generalization. An optimized Unet (O-Unet) architecture is designed for high performance prostate and tumor segmentation and subsequently, an Improved Gray-Level Co-occurrence Matrix (IGLCM) is developed for discriminative feature extraction from the segmented partitions. In the second stage of our framework, we explore three possibilities for classification: (i) Artificial Neural Network (ANN) using statistical features, (ii) Convolutional Neural Network (CNN) utilizing spatial features, and (iii) a hybrid ANN–CNN model that can take advantage of the strengths of both. Experimental results demonstrate that the task would be completed with increased diagnostic accuracy and robustness to MRI variances, and my clinical decision support would exceed existing CAD. The proposed CAD system performs accurate prostate and lesion segmentation followed by robust classification of prostate cancer using a hybrid ANN–CNN model.