Medical X-ray Image Classification and Encrypted Transmission System Based on Improved Residual Network
With the rapid development of X-ray imaging in the field of medical diagnosis,traditional manual judgment and analysis methods based on the doctor's experience cannot meet the efficiency requirements for diagnosing a large number of chest X-ray ima-ges.it is urgent to efficiently and accurately solve the classification of batch X-ray images.The residual network is introduced to clas-sify the batch X-ray images,and an encrypted transmission system is designed to effectively achieve the above problem.The X-ray im-ages are used to enhance the images based on the Markov random field,and then the ResNet50 with strong deep information mining a-bility is used as the backbone network to add the self-attention mechanism,and the continuously differentiable exponential linear units(CELU)activation function is used to optimize.The experiment is implemented on the Kaggle integrated dataset,the results show that the recall rate of classification is increased from 0.432 to 0.652 while ensuring the classification accuracy.Additionally,the im-age encryption algorithm based on logistic chaotic sequences is used to ensure the privacy of remote medical diagnosis,meeting the ap-plication requirements of actual remote medical scenarios.
X-ray images on chestResnet50Markov random fieldself-attentionCELUlogistic chaotic sequence
NETL
NSTL
万方数据
