Stroke was the cause of one out of every six deaths from cerebrovascular disease in 2020. A stroke occurs in the United States (US) every 40 seconds. Every 3.5 minutes, people die of a stroke. More than total 795,000 stroke cases occur yearly in the US. This study aims to detect the ischemic stroke lesion that occurs in the brain. The Ischemic Stroke Lesion Segmentation (ISLES) 2017 data set, which includes 82 Magnetic Resonance images of 43 patients, was used. The UNet, Attention UNet, Residual UNet, Attention Residual UNet, and Residual UNet++ segmentation networks were tested. Moreover, Cross Entropy, Dice, IoU, Tversky, Focal Tversky, and their compound forms were analyzed. The IoU loss function tested on Attention UNet achieved the best performance with the dice score of 0.766, the IoU score of 0.621, the sensitivity of 0.730, the specificity of 0.997, the precision of 0.805, and the accuracy of 0.993.

Çalışmada Alzheimer hastalığının analizi için sınıflandırma ve segmentasyon görevleri uygulanmıştır. Sınıflandırma görevinde transfer öğrenme kullanılarak 7 farklı model test edilmiştir. GoogLeNet modeli 0.9467 doğruluk, 0.9474 duyarlılık, 0.9811 özgüllük ve 0.9467 F1 skoru ile en iyi sınıflandırma performansını elde etmiştir. Segmentasyon görevinde, Alzheimer hastalığının segmentasyonu için U-Net mimari tasarımı kullanılmıştır. U-Net modeli 0.874 zar skoru, 0.776 IoU, 0.868 duyarlılık, 0.999 özgüllük, 0.879 kesinlik ve 0.999 doğruluk elde etmiştir. Pipeline oluşturmak için sınıflandırma ve segmentasyon modelleri birlikte kullanılmıştır. Sonuç olarak, bilgisayarlı görü destekli bir karar destek sistemi oluşturulmuştur.

The prevalence of brain tumor is quite high. Brain tumor causes critical diseases. Also, brain tumor causes a variety of symptoms in most people. This study aims to segmentation of the tumor in the brain. For this purpose, state-of-art architectures, such as UNet, Attention UNet, Residual UNet, Attention Residual UNet, Residual UNet++, Inception UNet, LinkNet, and SegNet were used for segmentation. 592 magnetic resonance (MR) images were utilized in the training and testing of segmentation architectures. In the comparative analysis, Attention UNet achieved the best predictive performance with a 0.886 dice score, 0.795 IoU score, 0.881 sensitivity, 0.993 specificity, 0.891 precision, and 0.986 accuracy.

Skull stripping has an important in neuroimaging workflow. Skull stripping is a time-consuming process in the Magnetic resonance imaging (MRI). For this reason, skull stripping and brain segmentation are aimed in this study. For the this purpose, the U-NET architecture design, which is one of the frequently used models in the field of medical image segmentation, was used. Also, different loss functions such as Cross Entropy (CE), Dice, IoU, Tversky, Focal Tversky and their compound forms were tested on U-Net architecture design. The compound loss function of CE and Dice loss functions achieved the best performace with the average dice score of 0.976, average IoU score of 0.964, sensitivity of 0.972, specificity of 0.985, precision of 0.960 and accuracy of 0.981. As a result, skull stripping was performed to facilitate the detection of brain diseases.

The pancreas is one of the vital organs in the human body. Early diagnosis of a disease in the pancreas is critical. In this way, the effects of pancreas diseases, especially pancreatic cancer on the person are decreased. With this purpose, artificial intelligence-assisted pancreatic cancer segmentation was performed for early diagnosis in this paper. For this aim, several state-of-the-art segmentation networks, UNet, LinkNet, SegNet, SQ-Net, DABNet, EDANet, and ESNet were used in this study. In the comparative analysis, the best segmentation performance has been achieved by SQ-Net. SQ-Net has achieved a 0.917 dice score, 0.847 IoU score, 0.920 sensitivity, 1.000 specificity, 0.914 precision, and 0.999 accuracy. Considering these results, an artificial intelligence-based decision support system was created in the study.