Texture Feature Analysis for Identification of AIS using Multi- Otsu’s Thresholding

Computing and characterizing Acute Ischemic Stroke (AIS) lesions remains a challenge. This technique precisely diagnoses the stroke lesion, allowing the radiologist to provide prompt and appropriate treatment with the goal of lowering the death rate due to brain stroke. The procedure is divided into three sections: preprocessing, multi-level thresholding, and feature extraction. The noise and skull part are removed in the pre-processing module, and then Multi Otsu's thresholding is applied. To identify the affected area, the image is divided into four regions. The affected region, which contains a stroke lesion, is statistically defined using a first-order histogram. The paper proposes an effective algorithm to support ischemic stroke lesion detection and feature extraction based on MRI images.