Through utilizing advanced deep learning technology, computers today are able to detect, locate, and image brain tumors, allowing a precise guide for both surgery and radiosurgery. Today, common tumors such as but not limited to brain metastases, meningioma, acoustic neuromas, and pituitary tumors are able to be detected through such technology.

Previously, technology is only capable of detecting one tumor or one disease. However, what we are aiming for is to be able to detect, locate, and image all sorts of tumors and diseases, including vascular diseases.
Currently, the model＇s mean dice score reaches 83.77% and the brain tumor＇s area reaches 67.56%

The main purpose of this project is to create a device that can firstly help radiologists to locate and illustrate brain tumors, and then allowing doctors and physicians to judge the results and determine the boundaries of the tumor for radiosurgical procedure.