Skull-stripping for Tumor-bearing Brain Images
This work addresses a specific challenge in medical imaging for clinicians and researchers dealing with brain tumors, but it appears incremental as it adapts skull-stripping to a niche domain.
The authors tackled the problem of skull-stripping for tumor-bearing brain images, where existing algorithms perform poorly, and developed a dedicated method to address this issue, though no concrete results or numbers are provided.
Skull-stripping separates the skull region of the head from the soft brain tissues. In many cases of brain image analysis, this is an essential preprocessing step in order to improve the final result. This is true for both registration and segmentation tasks. In fact, skull-stripping of magnetic resonance images (MRI) is a well-studied problem with numerous publications in recent years. Many different algorithms have been proposed, a summary and comparison of which can be found in [Fennema-Notestine, 2006]. Despite the abundance of approaches, we discovered that the algorithms which had been suggested so far, perform poorly when dealing with tumor-bearing brain images. This is mostly due to additional difficulties in separating the brain from the skull in this case, especially when the lesion is located very close to the skull border. Additionally, images acquired according to standard clinical protocols, often exhibit anisotropic resolution and only partial coverage, which further complicates the task. Therefore, we developed a method which is dedicated to skull-stripping for clinically acquired tumor-bearing brain images.