Classification of brain hemorrhage computed tomography (CT) images provides a better diagnostic implementation for emergency patients. Each brain CT image must be examined attentively by doctors. This situation takes time and sometimes leads to making errors. Hence, we aim to find the best algorithm owing to a requirement for automated brain hemorrhage detection. In this chapter, we utilized artificial intelligence for brain hemorrhage detection by using different machine learning and deep learning architectures. Although pretrained deep learning models achieve reasonable classification results, we utilize them for deep feature extraction by combining them with neighborhood components analysis (NCA) and classical machine learning techniques to achieve better performance. In these models, transfer learning models are utilized to extract features. These features are reduced to significant features with minimum loss by NCA. Eventually, we use different machine learning techniques to classify these significant features. Finally, experimental results reveal that the best-performing framework with a ResNet-18 feature extractor, NCA dimension reduction, and k-NN classifier achieves 96% accuracy with a brain hemorrhage CT dataset.