Domestic CPU retrieval method, system and device and computer readable medium

Abstract

The invention relates to the field of picture retrieval, and particularly provides a domestic CPU retrieval method which comprises the following steps: S1, a terminal uploads a picture to a background server; s2, the background server monitors the content uploaded by the government affair terminal through a monitoring program, and identifies characters on the official document picture; s3, non-standard character elements on the official document picture are converted into special textual characters to be stored; s4, performing word segmentation processing on the characters recognized in the step S2; s5, storing the recognized characters and the recognized special textual characters in a database; s6, establishing a full-text index database in the database, and performing mapping processing on the characters in the database; s7, inputting a retrieval keyword at the terminal to perform full-text retrieval; and S8, returning the retrieved official document picture and attachment list corresponding to the keyword. Compared with the prior art, the method has the advantages that the retrieval efficiency of workers is greatly improved, the retrieval recall ratio is increased, and the problem that official documents and characters in pictures cannot be retrieved is well solved.

Description

technical field [0001] The invention relates to the field of image retrieval, and specifically provides a domestic CPU retrieval method, system, device and computer-readable medium. Background technique [0002] With the continuous development of our country's economy, the number of picture files accounts for 1 / 3 of the total amount of e-government document data. If the full-text search of pictures cannot be supported, it will directly affect the recall rate of full-text search of all documents on the e-government management platform. In view of the actual situation of the existence of official documents, a considerable part exists in graphics formats such as TIF, JPG, BMP, etc., and general retrieval technology cannot retrieve graphics and text. If this problem cannot be solved, the system is incomplete and should be provided to users. The information is also incomplete. Most of the existing technical solutions are based on the search formula of structured data in the data...

AI Technical Summary

Problems solved by technology

[0003] The picture recognition function in the prior art is very simple, the recognition rate of the recognized text is relatively low, and the correct rate is relatively low. Most of the retrieval is based on database retrieval, the recall rate is low, and the integration rate with the e-government system is not high. The characteristics of official document source data unable to reflect

Images