Certificate information identification method, transposition, equipment and medium

Abstract

The invention discloses a certificate information identification method, transposition, equipment and a medium. The method comprises the following steps: obtaining an identification model generated bytraining a certificate sample including a counterfeit certificate picture based on a deep learning network, wherein the counterfeit certificate picture is a picture of an invalid certificate generated according to the content format of the real certificate picture; and identifying the certificate information in the to-be-identified certificate picture through the identification model. The methodcan relatively ensure the number of certificate samples used for deep learning network training, and further relatively ensure the identification accuracy of the identification model generated by training on the information in the certificate in the real application scene. In addition, the invention further provides a certificate information identification device and equipment and a medium, and the beneficial effects are the same as described above.

Description

technical field [0001] The invention relates to the field of deep learning, in particular to a certificate information identification method, transposition, equipment and media. Background technique [0002] Deep learning is a new field in machine learning research. Its motivation is to establish and simulate the neural network of the human brain for analysis and learning. It imitates the mechanism of the human brain to explain data, such as images, sounds and texts. The traditional deep learning model The generation of needs to be based on the training of a large number of samples. [0003] Currently, for banking, financial loans and other fields that involve entering user certificate information for business processing, when entering the information in the customer certificate into the system, it is necessary to accurately identify the information on it based on the scanned certificate picture, and then transform it into Text information, returned to business personnel fo...

AI Technical Summary

Problems solved by technology

In order to be able to realize the information recognition of the user's certificate by the device, it is often necessary to use the deep learning method to pre-train the user's certificate as a certificate sample for feature training, and then based on the recognition model generated by the training. recognition, but for the training of the recognition model, a large number of certificate pictures are required as samples. Since the certificate itself involves issues such as user privacy, the total number of certificate samples that can be collected and used for deep learning network feature training is relatively small, and it is difficult to ensure The recognition accuracy of the recognition model generated by training to the information in the certificate in the real application scenario

Images