Scholarlink

<p>The skull area of a head MRI possesses incomprehensible information that has aught significance in diagnosis of disease. Conversely, it can effectively degrade the performance of a computer-aided diagnostic (CAD) system and mislead the system into a false diagnosis. Therefore, it is coveted to isolate the skull from the input image before proceeding with the actual diagnosis process. For this purpose, the article proposes a simple algorithm based on digital image processing. The proposed method includes adaptive intensity slicing, adaptive morphology, and innovative ‘Effective Connected-Component Analysis’ operations. The fundamental concept of the proposed method is to separate the skull from the brain using adaptive morphology and then extract the brain component using ‘Effective Connected-Component Analysis’. The comparative analysis shows that the proposed method gives better Sensitivity and Specificity values (93.6 and 96.8) than the other image processing-based state-of-art methods. Moreover, the proposed method is very time efficient and does not require multiple image slices or training data sets.</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

A Simple Skull Stripping Approach from T1 Weighted MRI Using Adaptive Digital Image Processing Techniques

  • 期刊论文

  • 发表日期：

    2025年5月24日

查看全文
  • Tanmoy Kanti Halder,
  • Kanishka Sarkar,
  • Ardhendu Mandal,
  • Anil Tudu,
  • Bikramadittya Bagchi

摘要

The skull area of a head MRI possesses incomprehensible information that has aught significance in diagnosis of disease. Conversely, it can effectively degrade the performance of a computer-aided diagnostic (CAD) system and mislead the system into a false diagnosis. Therefore, it is coveted to isolate the skull from the input image before proceeding with the actual diagnosis process. For this purpose, the article proposes a simple algorithm based on digital image processing. The proposed method includes adaptive intensity slicing, adaptive morphology, and innovative ‘Effective Connected-Component Analysis’ operations. The fundamental concept of the proposed method is to separate the skull from the brain using adaptive morphology and then extract the brain component using ‘Effective Connected-Component Analysis’. The comparative analysis shows that the proposed method gives better Sensitivity and Specificity values (93.6 and 96.8) than the other image processing-based state-of-art methods. Moreover, the proposed method is very time efficient and does not require multiple image slices or training data sets.