Scholarlink

<p>Cone beam computed tomography (CBCT) scanning faces significant challenges in metal artifact correction. Medical image segmentation, particularly for CBCT dental scans, is complicated by the presence of metal artifacts that traditional approaches often struggle to address, resulting in degraded image quality and unreliable segmentation results. To tackle these challenges, this paper proposes a hybrid genetic algorithm based on synergy Thompson sampling hyper-heuristic (HGA-SyTSHH) for optimizing metal artifact reduction and enhancing CBCT image segmentation. Monte Carlo simulations are employed to generate realistic metal-corrupted CBCT projections, providing accurate training data. The proposed methodology integrates an iterative multilevel minimum cross entropy thresholding algorithm to enhance computational efficiency. Additionally, SyTSHH is utilized for optimal feature selection. Furthermore, the integration of SyTSHH with the HGA facilitates the dynamic selection of crossover and mutation operators, improving segmentation accuracy. The effectiveness of the proposed method is evaluated using a CBCT dental dataset, achieving a high segmentation accuracy of approximately 98.4%. This enhanced segmentation quality and reduced noise in metal-affected regions underscore the robustness and effectiveness of the proposed approach, making it well-suited for large-scale CBCT dental image segmentation tasks.</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

HGA-SyTSHH: CBCT Dental Segmentation Despite Metal Artifacts

  • 期刊论文

  • 发表日期：

    2025年11月6日

查看全文
  • T. S. Pradeep,
  • J. Arul Linsely

摘要

Cone beam computed tomography (CBCT) scanning faces significant challenges in metal artifact correction. Medical image segmentation, particularly for CBCT dental scans, is complicated by the presence of metal artifacts that traditional approaches often struggle to address, resulting in degraded image quality and unreliable segmentation results. To tackle these challenges, this paper proposes a hybrid genetic algorithm based on synergy Thompson sampling hyper-heuristic (HGA-SyTSHH) for optimizing metal artifact reduction and enhancing CBCT image segmentation. Monte Carlo simulations are employed to generate realistic metal-corrupted CBCT projections, providing accurate training data. The proposed methodology integrates an iterative multilevel minimum cross entropy thresholding algorithm to enhance computational efficiency. Additionally, SyTSHH is utilized for optimal feature selection. Furthermore, the integration of SyTSHH with the HGA facilitates the dynamic selection of crossover and mutation operators, improving segmentation accuracy. The effectiveness of the proposed method is evaluated using a CBCT dental dataset, achieving a high segmentation accuracy of approximately 98.4%. This enhanced segmentation quality and reduced noise in metal-affected regions underscore the robustness and effectiveness of the proposed approach, making it well-suited for large-scale CBCT dental image segmentation tasks.