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

Deep learning-based approximate video copy retrieval algorithm

A retrieval algorithm and deep learning technology, applied in the field of video approximate copy retrieval algorithm, can solve the problem of repeated uploading of similar videos, achieve the effect of small size, simple and efficient learning, and speed up the retrieval process

Active Publication Date: 2018-11-06
FUDAN UNIV
View PDF5 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a video approximate copy retrieval algorithm based on deep learning, which is applied to the approximate copy detection system of similar videos to solve the problem of repeated uploads of similar videos

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
  • Deep learning-based approximate video copy retrieval algorithm
  • Deep learning-based approximate video copy retrieval algorithm
  • Deep learning-based approximate video copy retrieval algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0044] The experiment selects 1000 public online videos in IACC.1.tv10.training in TRECVID as the offline video library, and the length of each video is between 3.6 and 4.1 minutes. After that, randomly select 100 videos from them, change their brightness, add subtitles, crop, scale, rotate (90 degrees and 180 degrees), add watermarks, draw frames and flip operations, plus the original 100 videos. This serves as ten sets of query videos.

[0045]Ten groups of query videos are input into the system for video approximate copy retrieval. The hyperparameter α in the calculation of video similarity takes a value of 0.1, and 12 groups of different confidence thresholds T are selected between 0 and 2.0 for experiments. After the candidate videos are obtained, they are respectively Calculate the similarity between them and the query video, and select the video with the highest similarity as the result output.

[0046] The comparison results of this experiment with the baseline method...

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 belongs to the technical field of image and video processing, and specifically relates to a deep learning-based approximate video copy retrieval algorithm. The algorithm is divided intotwo stages: an offline index construction stage and an online retrieval stage. In the offline index construction stage, feature values of sampling frames are extracted by using a deep convolutional neural network, and then indexes are established for all the video sampling frame sets in a video library by adoption of a k-d tree; and in the online retrieval stage, feature values of sampling framesof a queried video are extracted by adoption of the same method, similar candidate videos are searched from an index library by using an approximate nearest neighbor search method, similarities between all the candidate videos and the queried video are calculated finally, and the similarities are sorted from high to low to obtain an approximate copy retrieval result. According to the algorithm, the whole retrieval process can be greatly accelerated, and similarities between candidate videos and queried videos can be obtained for subsequent steps at the same time, so that the retrieval speed isfurther improved.

Description

technical field [0001] The invention belongs to the technical field of image and video processing, and in particular relates to a video approximate copy retrieval algorithm. Background technique [0002] With the great success of social networking sites and media, the number of videos increases rapidly, and similar or even identical videos are often uploaded repeatedly by different users. The video approximate copy detection technology used in the present invention can be considered as traditional content-based video approximate retrieval (videos should have similar visual content, but regardless of semantics) and semantic-based video retrieval (videos should have the same semantics, but Regardless of the visual content), it can effectively remove duplicate videos, thereby saving storage space and speeding up retrieval, and has broad application prospects. [0003] Most of the existing video approximate copy retrieval algorithms can be divided into three main steps: extract...

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): G06F17/30
Inventor 金城徐超吴渊张玥杰薛向阳
Owner FUDAN UNIV
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