© 2019, Springer Science+Business Media, LLC, part of Springer Nature.Detection of masses can be a challenging task for radiologists and physicians. Manual tumor diagnosis in the brain is sometimes a time consuming process and can be insufficient for fast and accurate detection and interpretation. This study introduces an improved interface-supported early diagnosis system to increase the speed and accuracy for supporting the traditional methods. The first stage in the system involves collecting information from the brain tissue, and assessing whether it is normal or abnormal through the processing of Magnetic Resonance Imaging (MRI) and Computerized Tomography (CT) images. The next stage involves gathering results from the image(s) after the single/multiple and volumetric and multiscale image analysis. The other stage involves Feature Extraction for some cases and making an interpretation about the abnormal Region of Interest (ROI) area via Deep Learning and conventional Artificial Intelligence methods is the last stage. The output of the system is mainly the name of the mass type which was introduced to the network. The results were obtained for totally 300 images for High-Grade Glioma (HGG), Low-Grade Glioma (LGG), Glioblastoma (GBM), Meningioma as well as Ischemic and Hemorrhagic stroke. For the cases, the DICE score was obtained as 0.927 and the normal/abnormal differentiation of the brain tissues was also achieved successfully. Finally, this system can give a chance to the doctors for supporting the results, speeding up the diagnosis process and also decreasing the rate of possible misdiagnosis.