Scholarlink

Purpose <p>The study introduces a SMAAHT-MIRRN-MTO model in deep learning focuses on improving diagnosis accuracy through X-rays for chest diseases. The system focuses on pneumonia and tuberculosis and COVID-19 diagnostics to enhance the reliability of medical diagnoses which improves healthcare quality.</p> Methods <p>The SMAAHT-MIRRN-MTO model includes three steps: preprocessing via Shape-aware Mesh Normal Filtering, segmentation using an Anatomy-Aware HoVer-Transformer, and classification with Multi-instance Riemannian Residual Neural Networks optimized by Mountaineering Team Optimization. The evaluation model uses accuracy as well as precision, recall and F1-score as performance metrics.</p> Results <p>A proposed model succeeded in analyzing chest X-rays effectively resulting in 99.93% accuracy along with 99.88% specificity and 99.93% sensitivity and F1-score performance that surpasses existing diagnostic systems in terms of precision and trust.</p> Conclusion <p>SMAAHT-MIRRN-MTO shows promise in the accurate detection of various chest diseases, showcasing deep learning’s potential to elevate diagnostic outcomes in medical imaging. The model’s high sensitivity and precision suggest practical applicability for clinical use, potentially reducing misdiagnosis rates.</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

Multi-instanceRiemannian residual neural network with mountaineering team-based chest disease detection using chest X-ray images

  • 期刊论文

  • 发表日期：

    2025年6月24日

查看全文
  • Raju Egala,
  • M. V. S. Sairam

摘要

Purpose

The study introduces a SMAAHT-MIRRN-MTO model in deep learning focuses on improving diagnosis accuracy through X-rays for chest diseases. The system focuses on pneumonia and tuberculosis and COVID-19 diagnostics to enhance the reliability of medical diagnoses which improves healthcare quality.

Methods

The SMAAHT-MIRRN-MTO model includes three steps: preprocessing via Shape-aware Mesh Normal Filtering, segmentation using an Anatomy-Aware HoVer-Transformer, and classification with Multi-instance Riemannian Residual Neural Networks optimized by Mountaineering Team Optimization. The evaluation model uses accuracy as well as precision, recall and F1-score as performance metrics.

Results

A proposed model succeeded in analyzing chest X-rays effectively resulting in 99.93% accuracy along with 99.88% specificity and 99.93% sensitivity and F1-score performance that surpasses existing diagnostic systems in terms of precision and trust.

Conclusion

SMAAHT-MIRRN-MTO shows promise in the accurate detection of various chest diseases, showcasing deep learning’s potential to elevate diagnostic outcomes in medical imaging. The model’s high sensitivity and precision suggest practical applicability for clinical use, potentially reducing misdiagnosis rates.