Classification and grading of white matter hyperintensity severity by using automated detection

Classification of WMH severity in MRI image can be considered as post processing stage or an additional procedure after segmentation process. This stage is significant for WMH detection system as it would further help to refine the image information and thus provide better interpretation for an expert analysis. In automated WMH detection and classification, the most common feature used for WMH classification is the WMH volume. As to dated, there is less research conducted particularly to classify the WMH severity based on intensity levels of WMH. This study is proposed to classify the severity of measured WMH based on its volume and intensity by automated approach. The MRI images sequences that were used for the processing are T2-WI and FLAIR images. In the automated WMH detection, FLAIR image was processed at the enhancement phase using the proposed enhancement technique whilst T2-WI image was processed at segmentation phase using two-tier segmentation technique. The potential WMHs obtained from FLAIR during enhancement phase and T2-WI during segmentation phase were mapped in final process to declare the most precise WMH regions in the MRI image. Finally, the detected WMHs were classified accordingly into three classes of severity namely as punctuate, focal and large.