The diagnosis of cerebral hemorrhage mainly relies on the imaging examination.The traditional manual way of the CT images analysis has many disadvantages,such as low identification efficiency,limited by doctors and equipment technology,and the difficult to achieve large-scale diagnosis.a new method of cerebral hemorrhage classification and recognition is proposed,based on convolutional neural network to overcome these difficulties.5 000 sets of brain CT images were collected,and the brain CT images were denoised,normalized and trimmed,and the LeNet-5 model of brain CT images was established for feature extraction and classification.The results show that the accuracy of the LeNet-5 model of cerebral hemorrhage recognition reached 95.8%on the test set,and it can provide more accurate and efficient diagnostic results,which can provide excellent support for clinical diagnosis and treatment.
维普
万方数据