Ischemic stroke is a leading cause of morbidity and mortality, requiring rapid and accurate detection for effective treatment. Segmentation of sulcal effacements in CT scan images is crucial for identifying acute strokes. In this paper, we propose a segmentation method that combines the Gray Level Co-occurrence Matrix (GLCM) and the Support Vector Machine (SVM) algorithm to enhance lesion identification efficiency. To achieve this, we developed a two-step approach integrating texturization and classification techniques to segment sulcal effacements in CT scan images. First, CT scan images are normalized and filtered to reduce noise and improve structural clarity, followed by segmentation to identify lesion areas. GLCM is then applied to each region of interest in the CT images to extract textural features such as contrast, correlation, energy, and homogeneity, which describe the properties of brain tissue. These features are used as input for the SVM algorithm, which classifies the effacements by determining the decision boundaries that separate different pixel groups in the feature space. The method is validated on a dataset annotated by experts, with segmentation performance evaluated using accuracy metrics. The results show that integrating GLCM and SVM improves the accuracy of sulcal effacement segmentation by up to 81% compared to traditional methods. The ROC curve of our model indicates good performance, demonstrating more precise detection of ischemic lesions.