Personalized image and subtitle generating method based on context sequence memory network

A technology of sequence generation and context, applied in neural learning methods, biological neural network models, electrical digital data processing, etc., can solve problems such as generation

Inactive Publication Date: 2017-10-13
SHENZHEN WEITESHI TECH
View PDF0 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] To solve the problem of generating personalized subtitles in images, the purpose of the present invention is to provide a method for generat

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Personalized image and subtitle generating method based on context sequence memory network
  • Personalized image and subtitle generating method based on context sequence memory network
  • Personalized image and subtitle generating method based on context sequence memory network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0047] It should be noted that, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present invention will be further described in detail below in conjunction with the drawings and specific embodiments.

[0048] figure 1 It is a system flowchart of the personalized image subtitle generation method based on the context sequence memory network of the present invention. It mainly includes database construction; contextual memory network; state-based sequence generation; network training.

[0049]Among them, the database construction utilizes the built-in application program interface of social media to generate a crawler device to capture the 10 most common tags in the posts published by social media;

[0050] (1) Crawled 3,455,021 original images with tags from posts shared by 17,813 users. These tags involve the 27 most common image categories in image sharing sites such as entertainmen...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a personalized image and subtitle generating method based on a context sequence memory network. The personalized image and subtitle generating method based on the context sequence memory network mainly comprises the following steps: construction of a database, construction of the context sequence memory network, sequence generation based on states and network training. According to the process, an image database is constructed by using an application interface in a social medium and a crawler device at first, redundant information is filtered, labels and subtitles are predicted by an image memory vector, an context memory vector and a character output vector, high frequency words are found out from a keyword dictionary, and therefore, personalized new words with the highest matching degree for the user are generated. Synchronous prediction of the labels and the subtitles can be processed, a frame of the context memory network is provided to solve the character matching problem, and meanwhile, the prediction precision of generated corresponding images and subtitles is improved.

Description

technical field [0001] The invention relates to the field of image tag prediction, in particular to a method for generating personalized image subtitles based on a context sequence memory network. Background technique [0002] It is one of the few emerging fields in recent years that the machine reads the image and judges the content of the image, and outputs the main content of the image with brief text to interact with humans. It mainly integrates image recognition, text retrieval and other technologies. Its purpose is to collect a large number of specific users. The image and its label and subtitle content, and predict the personalized subtitle of the upcoming new image with the characteristics of the user. The success of this method will greatly liberate manpower, especially in the fields of massive information processing and remote feedback. , such as the classification and labeling of interested users, social group studies, etc., have high development prospects. [00...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F17/30G06N3/04G06N3/08
CPCG06F16/5866G06N3/08G06N3/045
Inventor 夏春秋
Owner SHENZHEN WEITESHI TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap