A method and device for generating text based on pictures

Abstract

The present application provides a picture-based text generation method and device, wherein the method includes: determining at least one target vertical field, and obtaining the picture content information corresponding to the picture to be edited according to the target vertical field to which the picture to be edited belongs. The picture classification information corresponding to the picture to be edited; according to the picture content information corresponding to the picture to be edited and the picture classification information corresponding to the picture to be edited, generate the accurate picture information corresponding to the picture to be edited; according to the picture to be edited For the accurate information of the picture corresponding to the edited picture, the picture editing text corresponding to the picture to be edited is generated through the text generation model described in this application. This application obtains the accurate information of the picture corresponding to the picture to be edited according to the target vertical field, and edits the picture editing text with professional reference value in combination with the target vertical field, so as to assist the user to efficiently and accurately complete the professional document editing and writing work in the field.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, and in particular to an image-based text generation method, device, computing device, and computer-readable storage medium. Background technique [0002] When users create articles such as writing, such as news writing, they need to convert the picture information taken on the spot into text. This process requires users to have high observation ability and prior knowledge. If the picture information in the picture cannot be accurately identified , there will be errors in the user's understanding of the picture, and it will affect the accuracy of the article expression and writing efficiency. In the existing technology, the main ways to convert picture information into text include manual description and open-domain picture recognition and text conversion. Among them, manual description means that users manually edit the content of pictures by observing pictures, and open...

AI Technical Summary

Problems solved by technology

[0002] When users create articles such as writing, such as news writing, they need to convert the picture information taken on the spot into text. This process requires users to have high observation ability and prior knowledge. If the picture information in the picture cannot be accurately identified , there will be errors in the user's understanding of the picture, and it will affect the accuracy of the article expression and writing efficiency

In the existing technology, the main ways to convert picture information into text include manual description and open-domain picture recognition and text conversion. Among them, manual description means that users manually edit the content of pictures by observing pictures, and open-domain pictures Recognition and text conversion means to list the content in the picture for the user to use by detecting the target of the picture. However, the problem with the existing technology is that the manual description requires the user to have a lot of professional knowledge, and the accuracy is difficult to guarantee. The image recognition and text conversion in the open domain is only a simple list of targets, and cannot return accurate image information for a certain vertical field. At the same time, because the returned text content is a simple list of target names in the image, it cannot be directly used by users. less efficient

Images