Image processing concepts are used in the biomedical domain. Brain tumors are among the dreadful diseases. The primary brain tumors start with errors which are the mutations that take place in the part of DNA. This mutation makes the cells breed in a huge manner and makes the healthier cells die. The mass of the unhealthier cells is called tumor, or the unwanted cell growth in the tissues of the brain are called brain tumors. In this chapter, images from the positron emission tomography (PET) scan and MRI scan are fused as a one image, and from that image, neural network concepts are applied to detect the tumor. The main intention of this proposed approach is to segment and identify brain tumors in an automatic manner using image fusion with neural network concepts. Segmentation of brain images is needed to segment properly from other brain tissues. Perfect detection of size and position of the brain tumor plays an essential role in the identification of the tumor.