Computer Engineering Department, Fırat University, Elazığ, Turkey.
Med Hypotheses. 2020 Jun;139:109684. doi: 10.1016/j.mehy.2020.109684. Epub 2020 Mar 24.
Brain tumor is one of the dangerous and deadly cancer types seen in adults and children. Early and accurate diagnosis of brain tumor is important for the treatment process. It is an important step for specialists to detect the brain tumor using computer aided systems. These systems allow specialists to perform tumor detection more easily. However, mistakes made with traditional methods are also prevented. In this paper, it is aimed to diagnose the brain tumor using MRI images. CNN models, one of the deep learning networks, are used for the diagnosis process. Resnet50 architecture, one of the CNN models, is used as the base. The last 5 layers of the Resnet50 model have been removed and added 8 new layers. With this model, 97.2% accuracy value is obtained. Also, results are obtained with Alexnet, Resnet50, Densenet201, InceptionV3 and Googlenet models. Of all these models, the model developed with the highest performance has classified the brain tumor images. As a result, when analyzed in other studies in the literature, it is concluded that the developed method is effective and can be used in computer-aided systems to detect brain tumor.
脑肿瘤是成年人和儿童中常见的危险且致命的癌症类型之一。脑肿瘤的早期和准确诊断对于治疗过程非常重要。专家使用计算机辅助系统检测脑肿瘤是一个重要步骤。这些系统可以帮助专家更轻松地进行肿瘤检测。但是，也可以防止传统方法造成的错误。在本文中，旨在使用 MRI 图像诊断脑肿瘤。用于诊断过程的是深度学习网络之一的 CNN 模型。作为基础，使用了 CNN 模型之一的 Resnet50 架构。已经移除了 Resnet50 模型的最后 5 层，并添加了 8 个新层。通过该模型，获得了 97.2%的准确率值。此外，还使用了 Alexnet、Resnet50、Densenet201、InceptionV3 和 Googlenet 模型获得了结果。在所有这些模型中，使用性能最高的模型对脑肿瘤图像进行了分类。因此，在文献中的其他研究中进行分析后得出结论，所开发的方法是有效的，可以在计算机辅助系统中用于检测脑肿瘤。
