Separability of Acute Cerebral Infarction Lesions in CT Based Radiomics: Toward Artificial Intelligence-Assisted Diagnosis

Figure 2

The pipeline of our methodology included three steps: registration and candidate region acquisition, feature extraction and analysis, and classifiers establishment. Firstly, CT and MRI were input to obtain lesion regions and their symmetrical regions as candidate regions through registration. Then, features were extracted and calculated from candidate regions to capture useful features for auxiliary separating acute cerebral infarction. Finally, the classifiers were introduced to separate candidate region with selected features.