Method for discovering subject targets from video sequence

A technology of video sequences and topics, applied in the field of image processing, which can solve problems such as difficulty in summarizing

Active Publication Date: 2015-04-29
TSINGHUA UNIV
View PDF2 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The video contains a wealth of content, it is difficult for people to intuitively summarize

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
  • Method for discovering subject targets from video sequence
  • Method for discovering subject targets from video sequence
  • Method for discovering subject targets from video sequence

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0046] The method that the present invention proposes discovers subject target from video sequence, comprises the following steps:

[0047] (1) Train a classifier w and a gain coefficient v for selecting candidate regions containing subject objects in video images to be processed i and bias coefficient t i , the specific process is as follows:

[0048] (1-1) Take an image set containing known targets as a training sample, and adjust the image in the training sample to an image whose length and width are W and H respectively by interpolation method, where W,H∈{10,20, 40,80,160,320}, calculate the gradient norm of the adjusted image;

[0049] (1-2) Let T be all image windows containing known targets in the image, and S be any image window in the image, then the overlapping area o(t,s) of window t∈T and window s∈S is:

[0050] o ( t , s ) = t ...

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 relates to a method for discovering subject targets from a video sequence and belongs to the technical field of image processing. The method includes: training a classifier, a gain coefficient and a bias coefficient, and extracting two frames of images from to-be-processed video every second to obtain the video image sequence; selecting candidate subject targets from the image sequence, performing character representation on the candidate subject targets, and using a latent Dirichlet allocation model to discover the subject targets from the candidate subject targets. The method has the advantages that the candidate targets are obtained by a window method, so that target separation frequency caused by segmentation is avoided to a large extent; due to the fact that windows often contain the targets, the backgrounds in images are removed to a large extent during later model training, and the subjects obtained by latent Dirichlet allocation are accurate; a soft coding method allows character points to find accurate clustering centers during model iteration.

Description

technical field [0001] The invention relates to a method for discovering subject objects in video sequences, belonging to the technical field of image processing. Background technique [0002] In recent years, with the popularity of the Internet and television, people are faced with more and more video clips. The video contains a wealth of content, and it is difficult for people to intuitively summarize it. It is becoming more and more important to automatically find targets that match the subject of the video, so as to help people quickly understand the main content of the video. On the premise of retaining the main content of the original video, how to effectively extract the target in the video is the main difficulty of current research. [0003] A previous work is to find the target in a large number of pictures, segment each picture under different settings, and assume that these segmentations contain the target we want. Then use the Dirichlet distribution method to ...

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): G06K9/00G06K9/62
CPCG06V20/44G06V20/41G06F18/23213G06F18/2411
Inventor 刘华平刘云辉孙富春
Owner TSINGHUA UNIV
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