Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Label supplementing method and device

A label and filling technology, which is applied in the field of label filling methods and devices, can solve the problems of inability to fill labels in video, serious homogeneity, etc.

Inactive Publication Date: 2019-12-10
HUAWEI TECH CO LTD
View PDF6 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, if the video does not have a seed tag, such as a newly added video or an unpopular video, the technology cannot complete the tag for the video
In addition, since the labels are completed based on the similarity of the representation vectors, the labels of each video are very similar in semantics, and the homogeneity is serious.

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
  • Label supplementing method and device
  • Label supplementing method and device
  • Label supplementing method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0044] First, the terms involved in the embodiments of the present invention are explained, specifically as follows:

[0045] (1) Entity: In the embodiment of the present invention, it refers to an object or individual in a specific field, such as a video, a novel, a music, a commodity, etc.

[0046] (2) Representation vector: in the embodiment of the present invention, it refers to a vector representing a certain word. The basic idea is to map each word expressed in a specific language in a specific field into a fixed-length real number vector through the learning process (also known as the training process), and then form a vector space. In addition, a "distance measure" is introduced to calculate the lexical / semantic similarity of these words, where the "distance measure" can be considered as the distance between two vectors. For example, the representation vector of the word "cute" is (x 1 ,y 1 ,z 1 ), the representation vector of the word "sweet" is (x 2 ,y 2 ,z 2 ...

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 embodiment of the invention discloses a label supplementing method and device, relates to the technical field of multimedia, and aims to generate labels for videos with missing labels and avoid homogenization of different video labels to a certain extent. The method comprises the steps of obtaining a candidate label set of a first entity; wherein the candidate label set comprises a relationship representation vector and at least one label value representation vector; calculating the matching degree of each label value representation vector based on the relationship representation vector and the entity representation vector of the first entity; wherein the entity representation vector is a representation vector determined according to the identifier of the first entity; taking the labels corresponding to the first N label value representation vectors with higher matching degree with the entity representation vectors of the first entity as the labels of the first entity; wherein N isan integer greater than or equal to 1.

Description

technical field [0001] The embodiments of the present application relate to the field of multimedia technology, and in particular to a label complement method and device. Background technique [0002] Video operation websites can use video tags to mark video content to achieve fine-grained, multi-dimensional video indexing. In the existing technology, first, based on the existing video long comment and introduction, the representation vector of each tag in the existing video tags is trained, and the similarity between tags is calculated according to the representation vector, and similar tags are formed into similar tag clusters. For videos with relatively scarce tags, other tags in the similar tag cluster except the video's seed tag can be recommended to the video, so as to complete or expand the tags. [0003] However, if the video does not have a seed tag, such as a newly added video or an unpopular video, this technology cannot complete the tag for the video. In additi...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/78G06F16/71
Inventor 胡磊张旭
Owner HUAWEI TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com