In medical technology, Magnetic Resonance Imaging (MRI) has been used widely for detection of tumors and diagnosing various abnormalities in tissues. In scientific research, a major role has been played by the active development in the computerized segmentation of medical image. Based on fast decision making, the doctors can take required treatment easily. In the information technology, the segmentation of brain tumor is a key point. By analysing the radiation therapy treatment, tumor growth, computer-based surgery, developing the growth models of tumor, and treatment responses, the segmentation of brain tumor is motivated. For segmentation of brain tumor, a deep learning-based framework is presented. To achieve the robust performance through a majority rule, deep learning based semantic segmentation architecture is used for segmentation of tumor that can improve the performance effectively.