Scholarlink

<p>Ovarian cancer remains a highly dangerous type of gynecological cancer because patients receive late-stage diagnoses because none of the available screening methods manage to identify it early on. The detection of early ovarian cancer requires immediate detection because conventional diagnostic assessment methods using manual methods and traditional deep learning systems encounter numerous diagnostic limitations. Deep learning algorithms currently fail to detect multiple-length spatial patterns and time-based features as well as frequency patterns which reduces their diagnostic effectiveness. Switchable optimization systems are absent which produces diminishing returns in selecting features, creates computational processing setbacks and makes real-time medical application difficult to implement. This paper presents MedCross Recurrent Network (MCRNet) MCRNet with the FloraPotter-Gate (FPGate) Optimizer as novel techniques to boost ovarian cancer detection from medical images by providing a solution for these identified challenges. The MCRNet model achieves hierarchical spatial-temporal feature extraction through cross-modal recurrent learning with the purpose of maintaining important contextual information. A recursive optimization method uses dynamic parameter adjustments through a multi-objective loss function for stable convergence and optimized classification accuracy. Testing using benchmark ovarian cancer datasets confirms the strength of MCRNet-FPGate through its demonstrated 98.7% accuracy whereas standard CNNs and transformer-based models achieve inferior performance. The model achieves 98.3% F1-score alongside 97.5% sensitivity and 99.2% specificity which proves its competency in avoiding false detections and false negatives. The proposed framework delivers a streamlined computational process which leads to 27% decrease of complexity while providing deployment capacity. The results show that MCRNet-FPGate represents a leading approach for accurate automated ovarian cancer diagnosis which delivers efficient computation for its implementation.</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

MedCross recurrent network with adaptive gating optimization model for enhanced ovarian cancer detection and classification

  • 期刊论文

  • 发表日期：

    2026年2月28日

查看全文
  • S. Uma Maheswari,
  • S. Sudha

摘要

Ovarian cancer remains a highly dangerous type of gynecological cancer because patients receive late-stage diagnoses because none of the available screening methods manage to identify it early on. The detection of early ovarian cancer requires immediate detection because conventional diagnostic assessment methods using manual methods and traditional deep learning systems encounter numerous diagnostic limitations. Deep learning algorithms currently fail to detect multiple-length spatial patterns and time-based features as well as frequency patterns which reduces their diagnostic effectiveness. Switchable optimization systems are absent which produces diminishing returns in selecting features, creates computational processing setbacks and makes real-time medical application difficult to implement. This paper presents MedCross Recurrent Network (MCRNet) MCRNet with the FloraPotter-Gate (FPGate) Optimizer as novel techniques to boost ovarian cancer detection from medical images by providing a solution for these identified challenges. The MCRNet model achieves hierarchical spatial-temporal feature extraction through cross-modal recurrent learning with the purpose of maintaining important contextual information. A recursive optimization method uses dynamic parameter adjustments through a multi-objective loss function for stable convergence and optimized classification accuracy. Testing using benchmark ovarian cancer datasets confirms the strength of MCRNet-FPGate through its demonstrated 98.7% accuracy whereas standard CNNs and transformer-based models achieve inferior performance. The model achieves 98.3% F1-score alongside 97.5% sensitivity and 99.2% specificity which proves its competency in avoiding false detections and false negatives. The proposed framework delivers a streamlined computational process which leads to 27% decrease of complexity while providing deployment capacity. The results show that MCRNet-FPGate represents a leading approach for accurate automated ovarian cancer diagnosis which delivers efficient computation for its implementation.