Identify Brain Stroke in Severity and Symptoms in Sex Differences Using AIML

To aid professionals in measuring stroke lesions and developing treatment strategies, quick and accurate categorization of stroke lesions is extremely desirable. Due to stroke evolution, low contrast, and very varied grayscale distributions, segmenting stroke lesions is difficult. In order to achieve improved segmentation of stroke lesions, we offer a novel multi-task learning framework. In this method, we use supervised learning to improve the efficacy of the primary task avoiding increasing the time required to infer and model parameters in the stage of inference for the primary assignment (stroke lesion segmentation) and the auxiliary task (edge characterization of the brain lesion). For validation, a CT cerebral infarction data has been produced.