| abstract |
The invention relates to a tumor detection method, in particular to an intelligent detection method for hand tumors based on deep learning, and belongs to the technical field of medical image intelligent identification. An intelligent detection method for hand tumors based on deep learning, the method includes the following steps: (1) labeling magnetic resonance images of hand tumors; (2) preprocessing the labeling data, and enhancing the data set; (3) ) Build a fully convolutional neural network model, determine the parameters of the fully convolutional neural network model, input the data set into the fully convolutional neural network model, and use the loss function Perform training; where l(x) is the labeling category of pixel x, and W l(x) corresponds to the weight of category l(x). Use the deep learning method to perform feature learning from the data set, and obtain a trained fully convolutional neural network; (4) Input the image to be predicted into the intelligent detection model to obtain the prediction result. |