Method and device for extracting figure key frames from video file

A video file and video frame technology, applied in the field of image recognition, can solve the problems of difficult users, difficult to significantly improve the video click rate, unable to capture pictures, etc.

Inactive Publication Date: 2018-03-23
BEIJING QIHOO TECH CO LTD
View PDF5 Cites 27 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although the above method can attach related pictures to the video, its limitations are: the method of selecting pictures through manual compilation is very inefficient, and it is difficult to meet the huge daily video processing needs of the platform; The title of the video or the most relevant image to the content is as difficult to quickly attract users' attention as the method of directly using existing stills, so it is difficult to significantly increase the click-through rate of the video

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 and device for extracting figure key frames from video file
  • Method and device for extracting figure key frames from video file
  • Method and device for extracting figure key frames from video file

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0058] Embodiments of the present invention provide a method for extracting character key frames from video files, such as figure 1 As shown, the method includes:

[0059] Step S101, extract video frames in a video file at a preset period, and extract character tags in the title of the video file.

[0060] For the embodiment of the present invention, the video file is a video file whose cover or thumbnail is to be extracted with key frames of characters, which may be TV dramas, movies, entertainment information, trailers, etc. It is not limited in the embodiment of the present invention.

[0061] For the embodiment of the present invention, since the video is composed of a plurality of continuous picture groups, the image content of a group of continuous video frames does not change much, and only part of the video frames need to be extracted from the video file at a preset period to ensure The processing efficiency of extracting character key frames from video files.

[00...

Embodiment 2

[0078] Another possible implementation of the embodiment of the present invention, on the basis of the first embodiment, further includes the steps shown in the second embodiment, wherein,

[0079] Step S102 includes: performing face feature detection on each video frame through a preset face detection model, and extracting the detected face feature corresponding to at least one face from each video frame.

[0080] Wherein, the face detection model is a face detection model trained based on a convolutional neural network CNN or a deep neural network DNN.

[0081] For example, through 20 million material images, the convolutional neural network CNN or deep neural network DNN is trained with two categories including faces and excluding faces to obtain a face detection model.

[0082] For the embodiment of the present invention, face feature detection is performed on each video frame through the face detection model obtained based on convolutional neural network CNN or deep neura...

Embodiment 3

[0084] Another possible implementation of the embodiment of the present invention, on the basis of the first embodiment, also includes the steps shown in the third embodiment, wherein,

[0085] Also include before step S103:

[0086] Extracting sample data from a face database, the sample data being face images and corresponding person classifications;

[0087] The face classification model is obtained by training the sample data.

[0088] For the embodiment of the present invention, the sample data in the face database includes a preset number of person classifications and face images corresponding to each person classification.

[0089] For example, the sample data in the face database includes 1500 person categories, and each person category has 20 corresponding face images.

[0090] For the embodiment of the present invention, through the face images contained in the sample data in the face database, a preset number of person classifications are used to train the convolu...

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 present invention provides a method and a device for extracting figure key frames from a video file. The method comprises the steps of: extracting video frames from a video file in a preset cycle,and extracting figure tags from the title of the video file; extracting face features, corresponding to at least one face, included in each video frame from each video frame; performing classification of the face features based on a face classification model, and determining face classification corresponding to at least one face in each video frame; and determining video frames of faces only comprising the face classifications and being consistent to the figure tags from each video frame according to the figure classifications corresponding to at least one face in each video frame and the figure tags, and taking the video frames of faces as figure key frames. The method provided by the invention can achieve intelligent and rapid figure key frame extraction from the video file to meet hugedemands of video processing by a platform every day; and moreover, the extracted figure key frames have a high correlation degree to the video file titles or content to attract users' attention so asto substantially improve the video clicking rate.

Description

technical field [0001] The present invention relates to the technical field of image recognition, in particular, the present invention relates to a method and device for extracting character key frames from video files. Background technique [0002] With the development of Internet technology, watching videos on various platforms through the Internet has become a way for more and more people to entertain and pastime. In order to increase the click-through rate of a video, various platforms often use a picture of a person related to the title or content of the video as the cover or thumbnail of the video to attract more users to watch the video. [0003] As for how to obtain the cover picture or thumbnail of the video, the prior art usually implements it by manually intercepting the video frame or directly using the existing stills. Although the above method can attach related pictures to the video, its limitations are: the method of selecting pictures through manual compila...

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): G06K9/00
CPCG06V40/172G06V20/41G06V20/46
Inventor 余道明陈强黄君实李东亮张康
Owner BEIJING QIHOO TECH CO LTD
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