Application of machine learning in intelligent encryption for digital information of real-time image text under big data

In the context of big data, the exploration of the application effect of machine learning in intelligent encryption for real-time image text digital information aims to improve the privacy information security of people. Aiming at the problem of digital information leakage of real-time image text, the convolutional neural network is introduced and improved by adding a preprocessing module to form AlexNet, to encrypt the digital information of real-time image text. Besides, to take into account both the security effect and the real-time performance of the system, the image text is encrypted by the chaotic sequence generated by a one-dimensional chaotic system called Logistic-Sine and a multi-dimensional chaotic system named Lorenz. In this way, a real-time image text encryption model is constructed by combining the chaotic function and AlexNet. Finally, a simulation experiment is performed to analyze the performance of this model. The comparative analysis indicates that the recognition accuracy of feature extraction of image text by the intelligent encryption model reaches 94.37%, which is at least 3.05% higher than that of other neural network models by scholars in related fields. In the security analysis of image text encryption, the information entropy of pixel values at (0, 0) of the proposed model is close to the ideal value 8. Meanwhile, the value of the number of pixels change rate is generally more than 99.50%, and the value of the unified average changing intensity is generally more than 33.50%. This demonstrates that the model has good security in resisting attacks. Therefore, the constructed model can provide good security guarantee under the premise of ensuring the recognition accuracy, which can provide experimental basis for improving the security performance of real-time image text data in the future.