Scholarlink

<p>Long-term morbidity and death are linked to spinal trauma injuries. Tragic neurological impairment might result from spinal trauma that causes harm to the spinal cord. To avoid additional neurologic deterioration, spinal cord damage and mechanical imbalance must be diagnosed and treated promptly. Understanding the fundamental imaging methods is necessary for spine surgeons to diagnose, treat, and predict spinal cord injuries. Early detection and diagnosis of spinal cord injury can improve the patient's neurological condition. Thereby, the developed work concerns proposing a novel spinal injury detection model using CT images. Primarily, Wiener Filtering (WF) is employed for preprocessing the CT image. Then, segmentation is done with an Improved attention-based mask RCNN (IA-MRCNN) approach. The next phase is to extract features, during which, modified Local Gabor Transitional Pattern (LGTP), PHOG and color features are derived. For spinal injury detection, Bi-LSTM and modified Link Net (MLNet) classifiers are adopted to determine if a fracture exists or not. From the analysis, the proposed MLNet + Bi-LSTM attained a better accuracy of 0.956 for 90% training data when compared to existing methods.</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

Modified LinkNet with Bi-LSTM for Spinal Cord Segmentation and Injury Detection using CT Images

  • 期刊论文

  • 发表日期：

    2025年10月24日

查看全文
  • N. Chiranjeevi,
  • S. Shafiulla Basha

摘要

Long-term morbidity and death are linked to spinal trauma injuries. Tragic neurological impairment might result from spinal trauma that causes harm to the spinal cord. To avoid additional neurologic deterioration, spinal cord damage and mechanical imbalance must be diagnosed and treated promptly. Understanding the fundamental imaging methods is necessary for spine surgeons to diagnose, treat, and predict spinal cord injuries. Early detection and diagnosis of spinal cord injury can improve the patient's neurological condition. Thereby, the developed work concerns proposing a novel spinal injury detection model using CT images. Primarily, Wiener Filtering (WF) is employed for preprocessing the CT image. Then, segmentation is done with an Improved attention-based mask RCNN (IA-MRCNN) approach. The next phase is to extract features, during which, modified Local Gabor Transitional Pattern (LGTP), PHOG and color features are derived. For spinal injury detection, Bi-LSTM and modified Link Net (MLNet) classifiers are adopted to determine if a fracture exists or not. From the analysis, the proposed MLNet + Bi-LSTM attained a better accuracy of 0.956 for 90% training data when compared to existing methods.