Chinese and English mixed merchant store name identification method, system and device, and medium

Abstract

The invention discloses a Chinese and English mixed merchant store name identification method, system and device, and a medium, and the method comprises the steps: employing a Chinese merchant store name segmentation PixelLink model to segment a character region of a Chinese and English mixed merchant store name picture to be identified, and segmenting M segmented character regions; using an English merchant store name segmentation PixelLink model to segment a character area of the Chinese and English mixed merchant store name picture to be identified, and segmenting N segmented character areas; selecting a character area with the largest area from the M segmented character areas and the N segmented character areas respectively; splicing the two screened character regions with the maximumareas to obtain a finally segmented character region; performing deviation correction on the finally segmented character area by using an STN spatial transformation network; and performing character recognition on the corrected character area by using the Chinese and English mixed recognition CRNN model to obtain a Chinese recognition result and an English recognition result.

Description

technical field [0001] The present disclosure relates to a method, system, device and medium for identifying Chinese and English mixed merchant store names. Background technique [0002] The statements in this section merely mention background art related to the present disclosure and do not necessarily constitute prior art. [0003] At present, when some merchants cooperate with Meituan, UnionPay and other companies to access the Internet, the merchants will provide the photo information of the front of the door for the partner to review, and the partner needs to manually check whether the font information in the photo is consistent with the text information provided by the merchant; Recognizing font information in photos has the disadvantages of low efficiency, heavy workload, and high error rate, so it is extremely necessary to combine artificial intelligence technology for assistance. [0004] In the process of realizing the present disclosure, the inventors found that ...

AI Technical Summary

Problems solved by technology

[0008] 1. There are interference items such as angle tilt in the photos uploaded by users, which reduces the definition of text edges, which will reduce the accuracy of model recognition;

[0009] 2. Since most merchant stores contain both Chinese and English characters, it is difficult to use the same model to accurately identify Chinese and English characters due to the fact that Chinese characters are all square characters and English characters are relatively slender.

Images