Scholarlink

<p>Advanced computational methods like productive systems and deep neural models improve computerized division of tumors in MRI scans. These methods are essential for supporting planning of treatment and diagnosis because they increase efficiency and accuracy. However, there are drawbacks to current approaches, including poor accuracy and precision as well as high error rates. To overcome these challenges, automated multimodal brain tumor segmentation in MRI has been accomplished using a Patch-Based Pyramid Siamese 3D Convolutional Neural Network with Gorilla Troops Optimizer Algorithm (P-PS3DCNN-GTOA). This system utilizes data from three datasets: BraTS 2020, BraTS 2018–2019, and the Harvard datasets. The input data are pre-processed using Observability-Constrained Resampling-Free Cubature Kalman Filter (O-C-FCKF) method. Tumor segmentation is then performed using a Geometric Algebra Transformer (GAT), followed by feature extraction and Patch-Based Pyramid Siamese 3D Convolutional Neural Network (P-S3DCNN) for classification. The Gorilla Troops Optimizer Algorithm (GTOA) manages optimization process and successfully detects different forms of brain tumors and differentiates between normal and abnormal areas. Implemented in Python, suggested system achieves 99.9% accuracy, a 0.1% error percentage, and a 0.1-s process time (PT), demonstrating notable gains over current techniques. The outcomes show exceptional effectiveness and point to possibility of more developments in this area.</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

Patch-Based Pyramid Siamese 3D Convolutional Neural Network for Automated Multimodal Brain Tumor Segmentation in MRI Using Gorilla Troops Optimizer Algorithm

  • 期刊论文

  • 发表日期：

    2025年9月28日

查看全文
  • L. Mohana Kannan,
  • B. Pushpa,
  • R. Giri Prasad,
  • D. Samundeeswari

摘要

Advanced computational methods like productive systems and deep neural models improve computerized division of tumors in MRI scans. These methods are essential for supporting planning of treatment and diagnosis because they increase efficiency and accuracy. However, there are drawbacks to current approaches, including poor accuracy and precision as well as high error rates. To overcome these challenges, automated multimodal brain tumor segmentation in MRI has been accomplished using a Patch-Based Pyramid Siamese 3D Convolutional Neural Network with Gorilla Troops Optimizer Algorithm (P-PS3DCNN-GTOA). This system utilizes data from three datasets: BraTS 2020, BraTS 2018–2019, and the Harvard datasets. The input data are pre-processed using Observability-Constrained Resampling-Free Cubature Kalman Filter (O-C-FCKF) method. Tumor segmentation is then performed using a Geometric Algebra Transformer (GAT), followed by feature extraction and Patch-Based Pyramid Siamese 3D Convolutional Neural Network (P-S3DCNN) for classification. The Gorilla Troops Optimizer Algorithm (GTOA) manages optimization process and successfully detects different forms of brain tumors and differentiates between normal and abnormal areas. Implemented in Python, suggested system achieves 99.9% accuracy, a 0.1% error percentage, and a 0.1-s process time (PT), demonstrating notable gains over current techniques. The outcomes show exceptional effectiveness and point to possibility of more developments in this area.