Brain Tumor Medical Image Diagnostic Network Deployed on Low-Cost Embedded Devices

Early diagnosis of brain tumors is of great significance to patients, families, and society. Therefore, it is necessary to develop more accurate, fast, and reliable diagnostic methods for brain tumors. Deep learning based brain tumor diagnosis technology has great application prospects and is expected to play an important role in early diagnosis and treatment of brain tumors. This article proposes a lightweight deep learning model guided by attention mechanism. Training and testing were conducted on two publicly available brain tumor datasets, and compared with state-of-the-art methods. The experimental results show that the classification accuracy of this method is 99.8 %, which is higher than state-of-the-art methods. This has created a new state-of-the-art diagnostic method for brain tumors. The method proposed in this article is fast, robust, with low network computing complexity and low memory requirements. It can be deployed on low-cost embedded devices and has great application prospects in the medical field.