A method, device, device and client for generating text

Abstract

The embodiment of the present application discloses a method, device, device, and client for generating text. The method includes: receiving a first image, calculating image feature information of the first image, and determining the first image corresponding to the first image. A category information; in the first category, according to the image feature information of the first image, obtain the product information of a first preset number of candidate products corresponding to the first image; according to the product information of the candidate products information, calculating the first description feature information corresponding to the first image; according to the first description feature information and the pre-established mapping relationship between image description feature information and text keywords, determine the first description feature information corresponding to the first image a text. The embodiment of the present application can reduce the workload of the computer.

Description

technical field [0001] The present application relates to the technical field of computer data processing, in particular to a method, device, device and client for generating text. Background technique [0002] With the development of e-commerce and mobile Internet, more and more sellers use the client to publish their product information. At present, in order to further simplify the process for sellers to publish products, the client can automatically generate product text titles based on product pictures uploaded or taken by sellers. [0003] Existing methods for generating text titles generally include: extracting CNN (convolutional neural network) feature information from sample images, and extracting text keywords from the title information corresponding to the sample images; Words, using Recurrent Neural Network (RNN, recurrent neural network) or Long-Short TermMemory (LSTM, long-term short-term memory model) for training and learning, to obtain the mapping relationsh...

AI Technical Summary

Problems solved by technology

[0004] The inventors found that there are at least the following problems in the prior art: the CNN feature in the existing method of generating text titles is usually used to describe the image features of the product, but the product generally has a large number of categories, and it is impossible to distinguish products with similar image features Its categories, so the CNN feature information is difficult to accurately distinguish the image features of different categories of products under multiple categories

Therefore, when using the existing technology to produce text titles, it is necessary to establish a mapping relationship between image feature information and text keywords for each category. Therefore, the existing methods for generating text have poor category scalability and relatively heavy computer workload. Big

Images