A non-invasive and adaptive CAD system to detect brain tumor from T2-weighted MRIs using customized Otsu’s thresholding with prominent features and supervised learning
作者：
Highlights：
•
• A new non-invasive and adaptive CAD system based on brain MRI is developed for tumor detection. A case study is performed to identify Glioma tumors also.
• Otsu’s algorithm is customized for multi-level segmentation.
• The system is generalized for each view; Coronal, Sagittal and Axial. Also, the most prominent texture features are selected by a statistical parameter.
• Segmentation assessment is validated by statistical parameters, e.g., Dice Similarity Coefficient and Jaccard Index.
• The performance of the proposed CAD system is validated by Student’s t-test.
• T2-weighted MR images are considered for training and testing purposes. The system achieved 100% sensitivity (98.9% and 98.06% accuracy, respectively) on the two clinical datasets.
摘要
•A new non-invasive and adaptive CAD system based on brain MRI is developed for tumor detection. A case study is performed to identify Glioma tumors also.•Otsu’s algorithm is customized for multi-level segmentation.•The system is generalized for each view; Coronal, Sagittal and Axial. Also, the most prominent texture features are selected by a statistical parameter.•Segmentation assessment is validated by statistical parameters, e.g., Dice Similarity Coefficient and Jaccard Index.•The performance of the proposed CAD system is validated by Student’s t-test.•T2-weighted MR images are considered for training and testing purposes. The system achieved 100% sensitivity (98.9% and 98.06% accuracy, respectively) on the two clinical datasets.
论文关键词：Magnetic resonance imaging,Brain tumor,Multi-level thresholding,Computer aided diagnosis system
论文评审过程：Received 31 January 2016, Revised 19 May 2017, Accepted 19 May 2017, Available online 20 June 2017, Version of Record 7 November 2017.
论文官网地址：https://doi.org/10.1016/j.image.2017.05.013
