A medical document identification method based on an LSTM neural network

Abstract

The invention discloses a medical document identification method based on an LSTM neural network, and the method comprises the steps: 1) carrying out the preprocessing of a document image, and converting an image signal into a digital signal; 2) segmenting characters, normalizingf document images; 3) extracting character features, generating feature vectors; 4) performing document identification and classification.According to the method, an LSTM neural network is creatively adopted to recognize and classify the images, and the method has the advantages of being high in recognition speed, highin fault-tolerant capability, high in recognition rate, good in classification result and the like.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a medical bill recognition method based on LSTM neural network. Background technique [0002] In the insurance claims industry, claims documents including medical invoices, drug lists, medical records, inspection sheets, etc. are important claims basis. At present, for insurance companies, due to the need for data accumulation and regulatory requirements, the information collection of original documents is often very demanding. However, due to cost pressures, most insurance companies currently only collect invoice information through BPO, and other bill information It is often transformed into silent data, which cannot support the requirements of insurance companies for product design and automated control. The traditional BPO method mainly relies on manual entry, which requires manual classification of bills. The investment in personnel is huge, and data securi...

AI Technical Summary

Problems solved by technology

At present, for insurance companies, due to the need for data accumulation and regulatory requirements, the information collection of original documents is often very demanding. However, due to cost pressures, most insurance companies currently only collect invoice information through BPO, and other bill information It is often transformed into silent data, which cannot support the requirements of insurance companies for product design and automated control

The traditional BPO method mainly relies on manual entry, which requires manual classification of bills. The investment in personnel is huge, and data security management is complicated, and the overall efficiency is very low.

[0003] At present, there is no special identification method for insurance claims documents for the insurance industry. Nowadays, most of the financial bills are recognized, and the neural network pattern recognition method is used to segment and image the amount of the bill and the ID card number. processing and feature extraction, and on this basis, it was identified with an improved BP network

Images