Automatic brain abnormality detection in medical pictures requires great precision because it involves human lives. Also, medical facilities are actively seeking out computer assistance since it has the potential to improve human results in a field where the percentage of false negative and positive cases must be extremely low. Double-reading medical photos have been shown to improve the detection of aberrant regions. However, the expense of double-reading is substantial, which is why helpful software is in high demand at hospitals and other medical facilities. The chapter’s focus is on a computerised MR of the brain image digitization for the purposes of pre-processing features extraction and identifying brain abnormalities. Many methods for detecting normal and diseased tissues in the brain employ digitization as an intermediary step. The vast black background or the large change in contrast between background and foreground of MRI causes many pixels of brain portion to be incorrectly binarized, which is one of the fundamental challenges of MRI digitization. The proposed digitization uses the mean, variance, standard deviation, and entropy to establish a threshold value, after which the digitization problem can be solved with a non-gamut improvement. Extensive testing with multiple MRI types shows that the suggested digitization method produces accurate digitization with minimal human intervention. The results from this novel approach are compared to those from a more conventional one.