Abstarct: Abstract Today, brain tumors are one of the main causes of death worldwide. In 2012, according to statistics, the rate and death due to brain tumor was about 3.4% per 100,000 people and nervous system cancer was about 2.9% per 100,000 people around the world. Also, according to other studies, from 2004 to 2013 and in a period of 10 years, the rate of brain tumor and nervous system cancer has increased among people in terms of age. A very important point in dealing with this disease and the chances of its treatment is its correct and quick diagnosis. For this purpose, we need to use strong tools to diagnose and identify the disease. Applying artificial intelligence and using treatment methods for the brain is one of the very fast, strong and optimal methods in tumor diagnosis in the field of medical imaging, which is of interest to researchers in the field of machine vision. In this research, by using a privatized network, using unit socialization networks, brain tumor is identified and discovered in MRI images. In our activity, we used a module called adaptive gamma correction with the aim of enhancing the search features. Due to its features, this module makes better maps than the usual networks of the obtained images, thus improving the process of identifying and discovering medical tumors in the images. According to the existing basic structure of Gamant, by modifying the structure and architecture of this network and reducing the computational lines in this network, it showed the improvement of its computational complexity. The results of applying the proposed method, which was performed on the TCGA-LG dataset, showed the success of the proposed method due to the reduction of its computational complexity and the accuracy of the improvements compared to the basic architecture method and other existing methods.