Medical image information recognition method, device, and system based on multiple neural networks

Abstract

The present invention provides an information recognition method based on a multi-neural network, comprising: inputting a medical image as an input image into a first neural network model to obtain first identification information; inputting a medical image as an input image into a second neural network model to obtain Second identification information; when the first identification information and the second identification information are different, output an abnormal report and record the number of abnormal reports; adjust the first parameter of the first neural network model and the second neural network model according to the number of abnormal reports The second parameter of . The present invention superimposes at least two neural networks to reduce the misrecognition rate, adjusts the parameters of the neural network model according to the difference of multiple recognition results, and then performs image recognition on multiple neural network models, which reduces the error rate of recognition and corrects the error rate at the same time. The recognition rate is evaluated and the model is optimized to reduce the occurrence of medical report mismatches and avoid doctor-patient disputes.

Description

technical field [0001] The present invention relates to the technical field of information identification, in particular to a method, device and system for identifying medical image information based on a multi-neural network. Background technique [0002] The self-service printing system obtains patient films from imaging workstations such as DR, CT, MR, and PACS, and obtains patient information and patient reports such as patient names, mobile phone numbers, barcode numbers, medical card numbers, and ID card numbers from PACS reporting workstations, and passes specific rules. Match the patient's film to the report. Patients need to print out the corresponding film and report after swiping the barcode / swiping the doctor card / swiping the ID card. [0003] This report printing method increases the patient's waiting time for picking up the film, and if the patient's film information is directly sent to the patient or doctor, it is necessary to perform image recognition on the...

AI Technical Summary

Problems solved by technology

[0005] For the image recognition performed by OCR in the prior art, it is only considered to improve the recognition accuracy rate, but even if the recognition accuracy rate is improved, it is impossible to achieve 100% recognition accuracy, and it is impossible to know whether a recognition error has occurred during the recognition process, let alone correct The identification process is corrected

As a result, medical images are taken by mistake, and information errors such as names are usually not known until the doctor and / or patient hands, increasing doctor-patient disputes

When manual adjustments are introduced after human errors are discovered, efficiency will be reduced

Images