Research on Classification and Detection Mechanism of Brain Tumor based on SE-VGG-16

Brain tumors can affect both children and adults and afflict more than 510,000 individuals worldwide annually. To improve patients’ quality of life and survival rates, researchers investigated classification and detection methods of brain tumors. There are numerous classification models for brain tumors available right now, but most of them lack the ability to extract crucial information when extracting features, are very time-consuming, and are hampered by the extra regions of image features. This paper proposes a SE-VGG-16 model for brain tumor classification and detection. Firstly, threshold processing is used to the original data to make the images less disturbed by the extraneous influences, and the tumor regions is more obvious. Secondly, the SE module is added to each layer of VGG-16, the input image is extracted by SE module, and then the image is classified and output by full connection layer, and an evaluation model is constructed, the output classification results are compared with the real value. According to the experimental findings, brain tumor classification accuracy of SE-VGG-16 is 99.33 %, compared with the same type of detection mechanism, it has low classification error and high detection rate.