| Summary: | Low-field Magnetic Resonance Imaging (MRI) is vital in sensitive surgeries to allow real-time imaging in the operation theatre. A major concern in low-field MR images is the poor quality images and low signal-to-noise ratio (SNR) compared to those from conventional MRI scanners. For any successful application to be developed using such images, the image quality has to be improved. In this paper, pre-processing steps are applied to low-field MR brain images. Conceptually, histogram-based analysis shows that most of the low-field MR images consist of three peaks, where the first and second peaks summarizes the background and artifacts of the image, respectively, while the third peak is the region-of-interest (ROI). This paper gives some useful insights of steps that could be taken prior to brain segmentation. Promising results are reported.
|
|---|