CN113301383A - Video clip processing method, apparatus, and computer storage medium based on image feature analysis - Google Patents

Video clip processing method, apparatus, and computer storage medium based on image feature analysis Download PDF

Info

Publication number
CN113301383A
CN113301383A CN202110531926.3A CN202110531926A CN113301383A CN 113301383 A CN113301383 A CN 113301383A CN 202110531926 A CN202110531926 A CN 202110531926A CN 113301383 A CN113301383 A CN 113301383A
Authority
CN
China
Prior art keywords
comment
video
emotion
video content
words
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
CN202110531926.3A
Other languages
Chinese (zh)
Inventor
李刚
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan Dexin Yipin E Commerce Co ltd
Original Assignee
Wuhan Dexin Yipin E Commerce Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan Dexin Yipin E Commerce Co ltd filed Critical Wuhan Dexin Yipin E Commerce Co ltd
Priority to CN202110531926.3A priority Critical patent/CN113301383A/en
Publication of CN113301383A publication Critical patent/CN113301383A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/234—Processing of video elementary streams, e.g. splicing of video streams, manipulating MPEG-4 scene graphs
    • H04N21/23424—Processing of video elementary streams, e.g. splicing of video streams, manipulating MPEG-4 scene graphs involving splicing one content stream with another content stream, e.g. for inserting or substituting an advertisement
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G06F16/353—Clustering; Classification into predefined classes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/205—Parsing
    • G06F40/216—Parsing using statistical methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/234—Processing of video elementary streams, e.g. splicing of video streams, manipulating MPEG-4 scene graphs
    • H04N21/2343—Processing of video elementary streams, e.g. splicing of video streams, manipulating MPEG-4 scene graphs involving reformatting operations of video signals for distribution or compliance with end-user requests or end-user device requirements
    • H04N21/234336—Processing of video elementary streams, e.g. splicing of video streams, manipulating MPEG-4 scene graphs involving reformatting operations of video signals for distribution or compliance with end-user requests or end-user device requirements by media transcoding, e.g. video is transformed into a slideshow of still pictures or audio is converted into text
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/235—Processing of additional data, e.g. scrambling of additional data or processing content descriptors
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
    • H04N21/85—Assembly of content; Generation of multimedia applications
    • H04N21/858—Linking data to content, e.g. by linking an URL to a video object, by creating a hotspot

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Signal Processing (AREA)
  • Multimedia (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Probability & Statistics with Applications (AREA)
  • Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Television Signal Processing For Recording (AREA)

Abstract

Linking the movie and television video to a video playing website, extracting all comment barrages in the movie and television video playing process, extracting video content keywords and comment emotion words from the extracted comment barrages, further performing comprehensive analysis on the comment barrages to obtain highlight video content keywords corresponding to the movie and television video, simultaneously extracting audio information corresponding to the movie and television video, converting the audio information into text information, matching the highlight video content keywords with the converted text information, clipping the movie and television video according to the position of the successfully matched text words in the movie and television video to obtain highlight video segments corresponding to the movie and television video, clipping the highlight video segments, and greatly meeting the watching requirement of people who do not watch the movie and television on the highlight video segments in the movie and television video .

Description

Video clip processing method, apparatus, and computer storage medium based on image feature analysis
Technical Field
The invention belongs to the technical field of video clip processing, relates to a movie video clip technology, and particularly relates to a video clip processing method and device based on image characteristic analysis and a computer storage medium.
Background
Under the social environment of today, the great satisfaction of people's physical life directly promotes the increase of the demand of mental life, and film and television works as one of mental culture products play an important role in the process of meeting the demand of social mental life, so that a plurality of film and television videos are produced. However, with the continuous acceleration of the modern society rhythm, people may not have enough time to watch movie and television videos from beginning to end, and people hope to watch the wonderful part of the movie and television videos in a short time.
However, in the process of editing the movie and television video, how to screen the highlight segments in the movie and television video needs to be considered in an important way, and because the watching experience of each person on the movie and television video is different, the cognition of each person on the highlight segments of the movie and television video may be different, so that it is obviously unreasonable to screen the highlight segments in the movie and television video by simply using the cognition of each person on the highlight segments of the movie and television video, and the cognition of each person on the highlight segments of the movie and television video needs to be considered comprehensively, so that the highlight segments of the movie and television video are screened comprehensively.
Disclosure of Invention
In view of the above-mentioned needs, the present invention provides a video clip processing method, device and computer storage medium based on image feature analysis, wherein the movie and television video is linked to a video playing website, all comment banners in the playing process corresponding to the movie and television video are extracted, so that video content keywords and comment emotion words are extracted from the extracted comment banners, and then the extracted comment banners are comprehensively analyzed to obtain highlight video content keywords corresponding to the movie and television video, and at the same time, audio information corresponding to the movie and television video is extracted and converted into text information, so that the highlight video content keywords are matched with the converted text information, and the movie and television video is clipped according to successfully matched text words to obtain highlight video segments corresponding to the movie and television video, thereby realizing clipping of highlight segments in the movie and television video.
The purpose of the invention can be realized by the following technical scheme:
in a first aspect, the present invention provides a video clip processing method based on image feature analysis, comprising the steps of;
s1, segmenting a movie video: linking a movie video to be edited to a video playing website, acquiring the time length of the movie video, and dividing the time length of the movie video into a plurality of movie images according to a set video dividing frame number in the playing process of the movie video on the video playing website;
s2, comment barrage extraction: extracting the comment barrages existing on each segmented video image one by one, and summarizing the comment barrages to obtain all comment barrages in the corresponding playing process of the video;
s3, constructing a comment bullet screen set corresponding to the video content keywords: extracting video content keywords from each obtained comment barrage, classifying the comment barrages corresponding to the same video content keywords to obtain a comment barrage set corresponding to each video content keyword;
s4, obtaining a comprehensive comment barrage of the comment user: obtaining the comment user names corresponding to the comment barracks in the comment barrack set corresponding to the video content keywords, and combining the comment barrack contents corresponding to the same comment user in the comment barrack set corresponding to the video content keywords to obtain the comprehensive comment barracks of the comment users corresponding to the video content keywords;
s5, comment emotion type judgment of the comment user main body: extracting comment sentiment words from the comprehensive comment barrage of each comment user corresponding to each video content keyword, and judging the comment sentiment types corresponding to the extracted comment sentiment words, so as to analyze the comment sentiment words and obtain the main comment sentiment types corresponding to each comment user corresponding to each video content keyword;
s6, analyzing key words of the highlight video content: analyzing the main comment emotion types corresponding to the video content keywords and the comment users to obtain comprehensive comment emotion types corresponding to the video content keywords, comparing the comprehensive comment emotion types corresponding to the video content keywords with each other, and screening out the video content keywords corresponding to the comment emotion types, wherein the video content keywords are marked as wonderful video content keywords;
s7, converting video text information: extracting audio information corresponding to the movie and television videos, and converting the audio information into text information;
s8, video and video clip: matching the key words of the highlight video content with the converted text information, counting the number of successfully matched text words, and simultaneously respectively acquiring the corresponding in-point and out-point of each successfully matched text word in the movie video, so as to clip the movie video and obtain each video segment corresponding to the key words of the highlight video content;
s9, video clip merging: and merging the video segments corresponding to the key words of the wonderful video content to obtain video clip segments corresponding to the key words of the wonderful video content.
According to a preferred technical solution of the first aspect of the present invention, in S3, before extracting the video content keywords from the obtained comment barrages, the obtained comment barrages are subjected to deduplication processing, so as to obtain the deduplicated comment barrages.
According to a preferred technical solution of the first aspect of the present invention, in S3, the video content keywords are extracted from each obtained comment barrage, and a process of extracting the video content keywords is as follows:
h1, performing word segmentation on the content of each comment barrage to obtain each comment word, and performing part-of-speech tagging on each obtained comment word;
h2, comparing the obtained comment words with the stop word list one by one, if a part identical to a stop word in the stop word list exists in a certain comment word, filtering the comment word to obtain each reserved comment word, and recording the reserved comment word as a main comment word;
h3, according to the part of speech corresponding to the video content keywords, matching the part of speech corresponding to each reserved subject comment word with the part of speech corresponding to the video content keywords to obtain subject comment words consistent with the part of speech corresponding to the video content keywords, recording the subject comment words as target subject comment words, and further extracting the video content keywords from the target subject comment words.
According to a preferred technical scheme of the first aspect of the present invention, in S5, extracting comment sentiment words from the comprehensive comment barrage of each comment user corresponding to each video content keyword, and determining a comment sentiment type corresponding to the extracted comment sentiment words, a specific determination method is to compare the extracted comment sentiment words with a plurality of comment sentiment words corresponding to various comment sentiment types in a comment sentiment type information base, and if the comparison between the extracted comment sentiment words and any comment sentiment word corresponding to a certain comment sentiment type is successful, the comment sentiment type corresponding to the comment sentiment word is the comment sentiment type.
According to a preferred aspect of the first aspect of the present invention, the comment emotion types include a recognition type, a neutral type, and a derogation type.
According to a preferred technical solution of the first aspect of the present invention, the main comment emotion types corresponding to the comment users and corresponding to the video content keywords are obtained in S5, and the specific operation steps are as follows:
f1, counting the number of comment emotion words extracted from the comprehensive comment barrage of each comment user corresponding to each video content keyword;
f2, if only one comment sentiment word is extracted from the comprehensive comment barrage of a certain comment user, the main comment sentiment type corresponding to the comment user is the comment sentiment type corresponding to the comment sentiment word;
f3, if a plurality of comment emotion words are extracted from the comprehensive comment barrage of a comment user, comparing comment emotion types corresponding to the comment emotion words extracted from the comprehensive comment barrage of the comment user with each other, summarizing the comment emotion words corresponding to the same comment emotion types, counting the number of the same comment emotion types, and if only one comment emotion type is available, determining the main comment emotion type corresponding to the comment user as the same comment emotion type;
f4, if the same comment emotion types are multiple, numbering the same comment emotion types, respectively marking the comment emotion types as 1,2, a.i., a.n., n, counting the number of comment emotion words corresponding to the same comment emotion types, simultaneously numbering the comment emotion words corresponding to the same comment emotion types, sequentially marking the comment emotion words as 1,2, a.j.a.n., m, extracting emotion degree coefficients corresponding to the comment emotion types from a comment emotion type information base, respectively matching the comment emotion words corresponding to the same comment emotion types with the emotion degree coefficients corresponding to the comment emotion words corresponding to the comment emotion types, obtaining the emotion degree coefficients corresponding to the same comment emotion types, and counting the number of the comment emotion words corresponding to the same comment emotion types and the emotion degree coefficients corresponding to the comment emotion words according to the same comment emotion types The important weight coefficient corresponding to the type is calculated as
Figure BDA0003068213080000051
Figure BDA0003068213080000052
Is expressed as the important weight coefficient, k, corresponding to the ith same comment emotion typeiThe number, sigma, of comment emotion words corresponding to the ith same comment emotion typeij represents an emotion degree coefficient corresponding to the jth comment emotion word corresponding to the ith same comment emotion type;
f5, screening the same comment emotion type with the largest important weight coefficient from the important weight coefficients corresponding to the same comment emotion types as the main comment emotion type corresponding to the comment user.
According to a preferred technical solution of the first aspect of the present invention, the comprehensive comment emotion type corresponding to each video content keyword is obtained in S6, and the specific processing procedure includes the following steps:
g1, comparing the main comment emotion types corresponding to the comment users corresponding to the video content keywords, and classifying the comment users corresponding to the same main comment emotion types, wherein the same main comment emotion types are recorded as alternative comment emotion types;
g2, counting the number of the alternative comment emotion types corresponding to the video content keywords, wherein if only one alternative comment emotion type corresponds to a certain video content keyword, the comprehensive comment emotion type corresponding to the video content keyword is the alternative comment emotion type;
g3, if the alternative comment emotion types corresponding to the video content keyword are multiple, counting the number of comment users corresponding to each alternative comment emotion type, and further screening out the alternative comment emotion type with the largest number of comment users as the comprehensive comment emotion type corresponding to the video content keyword.
According to a preferred technical solution of the first aspect of the present invention, in the S8, the movie video is clipped, and the specific clipping method is to clip the movie video by using an in-point and an out-point of each successfully matched text word in the movie video as a clipping start position and a clipping end position of each successfully matched text word in the movie video.
In a second aspect, the present invention provides an apparatus, comprising a processor, and a memory and a network interface connected to the processor; the network interface is connected with a nonvolatile memory in the server; the processor calls the computer program from the nonvolatile memory through the network interface during running, and runs the computer program through the memory to execute the video clip processing method based on the image characteristic analysis.
In a third aspect, the present invention provides a computer storage medium, where a computer program is burned in the computer storage medium, and when the computer program runs in a memory of a server, the video clip processing method based on image feature analysis according to the present invention is implemented.
Based on any one of the above aspects, the invention has the following beneficial effects:
(1) the invention links the movie and television video to a video playing website, extracts all comment barrages in the corresponding playing process of the movie and television video, extracts the video content keywords and comment sentiment words from the extracted comment barrages, further performs comprehensive analysis on the comment barrages to obtain the highlight video content keywords corresponding to the movie and television video, extracts the audio information corresponding to the movie and television video at the same time, converts the audio information into text information, matches the highlight video content keywords with the converted text information, clips the movie and television video according to the successfully matched text words to obtain the highlight video corresponding to the movie and television video, realizes clipping of highlight segments of the movie and television video, greatly meets the watching requirements of people who do not watch the movie and television on the highlight video segments in the movie and television video, and screens the highlight video segments integrate the cognition of the people who watch the movie and television on the highlight segments, the method can be comprehensively practical, improves the reliability of the screening result, and further enhances the watching experience of people.
(2) According to the method, the extracted comment barrage is subjected to duplicate removal before the extracted comment barrage is subjected to video content keyword extraction and comment emotion word extraction, so that the problems of interference and useless work caused by repeated comment barrages in video content keyword extraction and comment emotion word extraction are solved, and the screening efficiency is improved for screening the corresponding highlight video segment of the movie video in the later period.
(3) According to the method, the movie and television videos are linked to the video playing website, all comment barrages in the corresponding playing process of the movie and television videos are extracted, the comment barrages serve as the cognition of each viewer who has watched the movie and television videos on the highlight segments in the movie and television videos, the obtaining mode has the advantages of being strong in practicability and strong in operability, the extracted comment barrages can truly reflect the watching experience of the viewer on the movie and television videos, and reliable and real reference basis is provided for screening the highlight segments corresponding to the movie and television videos.
Drawings
The invention is further illustrated by means of the attached drawings, but the embodiments in the drawings do not constitute any limitation to the invention, and for a person skilled in the art, other drawings can be obtained on the basis of the following drawings without inventive effort.
FIG. 1 is a flow chart of the method steps of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, in a first aspect, the present invention provides a video clip processing method based on image feature analysis, including the following steps;
s1, segmenting a movie video: linking a movie video to be edited to a video playing website, acquiring the time length of the movie video, and dividing the time length of the movie video into a plurality of movie images according to a set video dividing frame number in the playing process of the movie video on the video playing website;
in the embodiment, the movie and television video is divided into a plurality of movie and television images in the playing process of the video playing website, so that a foundation is laid for analyzing the movie and television images and extracting the user comment barrages in the movie and television images;
s2, comment barrage extraction: extracting the comment barrages existing on each segmented video image one by one, and summarizing the comment barrages to obtain all comment barrages in the corresponding playing process of the video;
s3, constructing a comment bullet screen set corresponding to the video content keywords: carrying out duplication removal processing on all the obtained comment barrages, and respectively extracting video content keywords from each duplicated comment barrage, wherein the process of extracting the video content keywords is as follows:
h1, performing word segmentation on the content of each comment barrage to obtain each comment word, and performing part-of-speech tagging on each obtained comment word;
h2, comparing the obtained comment words with the stop word list one by one, if a part identical to a stop word in the stop word list exists in a certain comment word, filtering the comment word to obtain each reserved comment word, and recording the reserved comment word as a main comment word;
h3, matching the part of speech corresponding to each reserved subject comment word with the part of speech corresponding to the video content keyword according to the part of speech corresponding to the video content keyword to obtain subject comment words consistent with the part of speech corresponding to the video content keyword, recording the subject comment words as target subject comment words, and further extracting the video content keyword from the target subject comment words;
classifying the comment barrages corresponding to the same video content keywords to obtain a comment barrage set corresponding to each video content keyword;
in the embodiment, the extracted comment barrage is subjected to duplicate removal processing, so that the problems of interference and useless work caused by extraction of the repeated comment barrage on the video content key words and extraction of the comment emotional words are avoided, and the screening efficiency is improved for screening of the corresponding wonderful video segment of the movie video in the later period;
s4, obtaining a comprehensive comment barrage of the comment user: obtaining the comment user names corresponding to the comment barracks in the comment barrack set corresponding to the video content keywords, and combining the comment barrack contents corresponding to the same comment user in the comment barrack set corresponding to the video content keywords to obtain the comprehensive comment barracks of the comment users corresponding to the video content keywords;
the comment barrage content corresponding to the same comment user is merged, so that convenience is brought to the subsequent targeted analysis of the comment barrage content corresponding to the same user;
in the embodiment, the movie and television video is linked to the video playing website, all comment barrages in the corresponding playing process of the movie and television video are extracted, and then the comment barrages are used as the cognition of each viewer who has watched the movie and television video on the highlight segments in the movie and television video, so that the obtaining mode has the characteristics of strong practicability and strong operability, the extracted comment barrages can truly reflect the watching experience of the viewer on the movie and television video, and reliable and real reference basis is provided for screening the highlight segments corresponding to the movie and television video;
s5, comment emotion type judgment of the comment user main body: extracting comment sentiment words of the comprehensive comment barrage of each comment user corresponding to each video content keyword, wherein the extraction process is as follows:
z1, carrying out word segmentation and stop word removal on the comprehensive comment barrage content of each comment user corresponding to each video content keyword to obtain each comment phrase, and carrying out duplication removal processing on each comment phrase to obtain a duplicate-removed comment phrase;
z2, performing part-of-speech tagging on each de-duplicated comment phrase, acquiring part-of-speech corresponding to the comment emotional words, and comparing the part-of-speech corresponding to each de-duplicated comment phrase with the part-of-speech corresponding to the comment emotional words so as to screen out comment phrases consistent with the part-of-speech of the comment emotional words, wherein the comment phrases are marked as candidate phrases;
z3, acquiring common words corresponding to the comment emotion words, matching each candidate phrase with the common words corresponding to the comment emotion words, and if matching of a word in a certain candidate phrase with a common word corresponding to the comment emotion words is successful, taking the candidate phrase as the comment emotion words;
judging a corresponding comment emotion type for the extracted comment emotion words, wherein the specific judgment method comprises the steps of comparing the extracted comment emotion words with a plurality of comment emotion words corresponding to various comment emotion types in a comment emotion type information base, and if the extracted comment emotion words are successfully compared with any comment emotion word corresponding to a certain comment emotion type, the comment emotion type corresponding to the comment emotion words is the comment emotion type, wherein the comment emotion types comprise a positive type, a neutral type and a negative type;
the comment emotion type information base is used for storing a plurality of comment emotion words corresponding to various comment emotion types and emotion degree coefficients corresponding to the comment emotion words;
at the moment, the comment emotion types corresponding to the comment emotion words extracted from the comprehensive comment barrage of each comment user corresponding to each video content keyword are analyzed to obtain the main comment emotion types corresponding to each comment user corresponding to each video content keyword, and the specific operation steps are as follows:
f1, counting the number of comment emotion words extracted from the comprehensive comment barrage of each comment user corresponding to each video content keyword;
f2, if only one comment sentiment word is extracted from the comprehensive comment barrage of a certain comment user, the main comment sentiment type corresponding to the comment user is the comment sentiment type corresponding to the comment sentiment word;
f3, if a plurality of comment emotion words are extracted from the comprehensive comment barrage of a comment user, comparing comment emotion types corresponding to the comment emotion words extracted from the comprehensive comment barrage of the comment user with each other, summarizing the comment emotion words corresponding to the same comment emotion types, counting the number of the same comment emotion types, and if only one comment emotion type is available, determining the main comment emotion type corresponding to the comment user as the same comment emotion type;
f4, if the same comment emotion types are multiple, numbering the same comment emotion types, marking the same comment emotion types as 1,2, aMatching the emotion degree coefficients corresponding to the comment emotion words to obtain the emotion degree coefficients corresponding to the comment emotion words of the same comment emotion types, and counting the important weight coefficients corresponding to the same comment emotion types according to the number of the comment emotion words corresponding to the same comment emotion types and the emotion degree coefficients corresponding to the comment emotion words, wherein the calculation formula is
Figure BDA0003068213080000111
Figure BDA0003068213080000112
Is expressed as the important weight coefficient, k, corresponding to the ith same comment emotion typeiThe number, sigma, of comment emotion words corresponding to the ith same comment emotion typeij represents an emotion degree coefficient corresponding to the jth comment emotion word corresponding to the ith same comment emotion type;
f5, screening the same comment emotion type with the largest important weight coefficient from the important weight coefficients corresponding to the same comment emotion types as the main comment emotion type corresponding to the comment user;
in the method, in the process of analyzing the main comment emotion types corresponding to the comment users corresponding to the video content keywords, various conditions are comprehensively considered, and a targeted processing mode is provided for various conditions, so that the main comment emotion type analysis result is more accurate and reliable;
s6, analyzing key words of the highlight video content: analyzing the main comment emotion types corresponding to the comment users and corresponding to the video content keywords to obtain the comprehensive comment emotion types corresponding to the video content keywords, wherein the specific processing process comprises the following steps:
g1, comparing the main comment emotion types corresponding to the comment users corresponding to the video content keywords, and classifying the comment users corresponding to the same main comment emotion types, wherein the same main comment emotion types are recorded as alternative comment emotion types;
g2, counting the number of the alternative comment emotion types corresponding to the video content keywords, wherein if only one alternative comment emotion type corresponds to a certain video content keyword, the comprehensive comment emotion type corresponding to the video content keyword is the alternative comment emotion type;
g3, if the alternative comment emotion types corresponding to the certain video content keyword are multiple, counting the number of comment users corresponding to each alternative comment emotion type, and further screening out the alternative comment emotion type with the largest number of comment users as the comprehensive comment emotion type corresponding to the video content keyword;
at the moment, the comprehensive comment emotion types corresponding to all the video content keywords are mutually compared, and the video content keywords corresponding to the comment emotion types are screened out from the comprehensive comment emotion types and are marked as wonderful video content keywords;
s7, converting video text information: extracting audio information corresponding to the movie and television videos, and converting the audio information into text information;
s8, video and video clip: matching key words of the highlight video content with the converted text information, and counting the number of successfully matched text words, wherein the corresponding positions of the successfully matched text words in the movie video are highlights corresponding to the movie video, and at the moment, respectively acquiring corresponding in-points and out-points of the successfully matched text words in the movie video, so as to clip the movie video;
s9, video clip merging: and numbering the video segments corresponding to the key words of the wonderful video content according to the sequence of the video segments in the time stamps corresponding to the movie and television videos, and combining the video segments corresponding to the key words of the wonderful video content according to the numbering sequence to obtain the video clip segments corresponding to the key words of the wonderful video content.
The invention links the movie and television video to a video playing website, extracts all comment barrages in the corresponding playing process of the movie and television video, extracts the video content keywords and comment sentiment words from the extracted comment barrages, further performs comprehensive analysis on the comment barrages to obtain the highlight video content keywords corresponding to the movie and television video, extracts the audio information corresponding to the movie and television video at the same time, converts the audio information into text information, matches the highlight video content keywords with the converted text information, clips the movie and television video according to the successfully matched text words to obtain the highlight video segment corresponding to the movie and television video, realizes clipping of the highlight video segment, greatly meets the watching requirements of people not watching the movie and television on the highlight video segment in the movie and television video, and filters the highlight video segment comprehensively realizes the cognition of people watching the movie and television on the highlight segment, the method can be comprehensively practical, improves the reliability of the screening result, and further enhances the watching experience of people.
In a second aspect, the present invention provides an apparatus, comprising a processor, and a memory and a network interface connected to the processor; the network interface is connected with a nonvolatile memory in the server; the processor calls the computer program from the nonvolatile memory through the network interface during running, and runs the computer program through the memory to execute the video clip processing method based on the image characteristic analysis.
In a third aspect, the present invention provides a computer storage medium, where a computer program is burned in the computer storage medium, and when the computer program runs in a memory of a server, the video clip processing method based on image feature analysis according to the present invention is implemented.
The foregoing is merely exemplary and illustrative of the present invention and various modifications, additions and substitutions may be made by those skilled in the art to the specific embodiments described without departing from the scope of the invention as defined in the following claims.

Claims (10)

1. The video clip processing method based on image feature analysis is characterized by comprising the following steps of:
s1, segmenting a movie video: linking a movie video to be edited to a video playing website, acquiring the time length of the movie video, and dividing the time length of the movie video into a plurality of movie images according to a set video dividing frame number in the playing process of the movie video on the video playing website;
s2, comment barrage extraction: extracting the comment barrages existing on each segmented video image one by one, and summarizing the comment barrages to obtain all comment barrages in the corresponding playing process of the video;
s3, constructing a comment bullet screen set corresponding to the video content keywords: extracting video content keywords from each obtained comment barrage, classifying the comment barrages corresponding to the same video content keywords to obtain a comment barrage set corresponding to each video content keyword;
s4, obtaining a comprehensive comment barrage of the comment user: obtaining the comment user names corresponding to the comment barracks in the comment barrack set corresponding to the video content keywords, and combining the comment barrack contents corresponding to the same comment user in the comment barrack set corresponding to the video content keywords to obtain the comprehensive comment barracks of the comment users corresponding to the video content keywords;
s5, comment emotion type judgment of the comment user main body: extracting comment sentiment words from the comprehensive comment barrage of each comment user corresponding to each video content keyword, and judging the comment sentiment types corresponding to the extracted comment sentiment words, so as to analyze the comment sentiment words and obtain the main comment sentiment types corresponding to each comment user corresponding to each video content keyword;
s6, analyzing key words of the highlight video content: analyzing the main comment emotion types corresponding to the video content keywords and the comment users to obtain comprehensive comment emotion types corresponding to the video content keywords, comparing the comprehensive comment emotion types corresponding to the video content keywords with each other, and screening out the video content keywords corresponding to the comment emotion types, wherein the video content keywords are marked as wonderful video content keywords;
s7, converting video text information: extracting audio information corresponding to the movie and television videos, and converting the audio information into text information;
s8, video and video clip: matching the key words of the highlight video content with the converted text information, counting the number of successfully matched text words, and simultaneously respectively acquiring the corresponding in-point and out-point of each successfully matched text word in the movie video, so as to clip the movie video and obtain each video segment corresponding to the key words of the highlight video content;
s9, video clip merging: and merging the video segments corresponding to the key words of the wonderful video content to obtain video clip segments corresponding to the key words of the wonderful video content.
2. The image feature analysis based video clip processing method of claim 1, wherein: and in the step S3, before extracting the video content keywords from the obtained comment barrages, performing deduplication processing on the obtained comment barrages to obtain the deduplicated comment barrages.
3. The image feature analysis based video clip processing method of claim 1, wherein: and S3, extracting the video content keywords from each obtained comment barrage, wherein the process of extracting the video content keywords is as follows:
h1, performing word segmentation on the content of each comment barrage to obtain each comment word, and performing part-of-speech tagging on each obtained comment word;
h2, comparing the obtained comment words with the stop word list one by one, if a part identical to a stop word in the stop word list exists in a certain comment word, filtering the comment word to obtain each reserved comment word, and recording the reserved comment word as a main comment word;
h3, according to the part of speech corresponding to the video content keywords, matching the part of speech corresponding to each reserved subject comment word with the part of speech corresponding to the video content keywords to obtain subject comment words consistent with the part of speech corresponding to the video content keywords, recording the subject comment words as target subject comment words, and further extracting the video content keywords from the target subject comment words.
4. The image feature analysis based video clip processing method of claim 1, wherein: in the step S5, extracting comment sentiment words from the comprehensive comment barrage of each comment user corresponding to each video content keyword, and determining a comment sentiment type corresponding to the extracted comment sentiment words, in which the specific determination method is to compare the extracted comment sentiment words with a plurality of comment sentiment words corresponding to each comment sentiment type in a comment sentiment type information base, and if the comparison of the extracted comment sentiment words with any comment sentiment word corresponding to a certain comment sentiment type is successful, the comment sentiment type corresponding to the comment sentiment words is the comment sentiment type.
5. The image feature analysis based video clip processing method of claim 4, wherein: the comment emotion types include a comment type, a neutral type, and a derogation type.
6. The image feature analysis based video clip processing method of claim 1, wherein: obtaining the main comment emotion types corresponding to the comment users and corresponding to the video content keywords in the step S5, wherein the specific operation steps are as follows:
f1, counting the number of comment emotion words extracted from the comprehensive comment barrage of each comment user corresponding to each video content keyword;
f2, if only one comment sentiment word is extracted from the comprehensive comment barrage of a certain comment user, the main comment sentiment type corresponding to the comment user is the comment sentiment type corresponding to the comment sentiment word;
f3, if a plurality of comment emotion words are extracted from the comprehensive comment barrage of a comment user, comparing comment emotion types corresponding to the comment emotion words extracted from the comprehensive comment barrage of the comment user with each other, summarizing the comment emotion words corresponding to the same comment emotion types, counting the number of the same comment emotion types, and if only one comment emotion type is available, determining the main comment emotion type corresponding to the comment user as the same comment emotion type;
f4, if the same comment emotion types are multiple, numbering the same comment emotion types, respectively marking the comment emotion types as 1,2, a.i., a.n., n, counting the number of comment emotion words corresponding to the same comment emotion types, simultaneously numbering the comment emotion words corresponding to the same comment emotion types, sequentially marking the comment emotion words as 1,2, a.j.a.n., m, extracting emotion degree coefficients corresponding to the comment emotion types from a comment emotion type information base, respectively matching the comment emotion words corresponding to the same comment emotion types with the emotion degree coefficients corresponding to the comment emotion words corresponding to the comment emotion types, obtaining the emotion degree coefficients corresponding to the same comment emotion types, and counting the number of the comment emotion words corresponding to the same comment emotion types and the emotion degree coefficients corresponding to the comment emotion words according to the same comment emotion types The important weight coefficient corresponding to the type is calculated as
Figure FDA0003068213070000041
Figure FDA0003068213070000042
Is expressed as the important weight coefficient, k, corresponding to the ith same comment emotion typeiThe number, sigma, of comment emotion words corresponding to the ith same comment emotion typeij represents an emotion degree coefficient corresponding to the jth comment emotion word corresponding to the ith same comment emotion type;
f5, screening the same comment emotion type with the largest important weight coefficient from the important weight coefficients corresponding to the same comment emotion types as the main comment emotion type corresponding to the comment user.
7. The image feature analysis based video clip processing method of claim 1, wherein: the comprehensive comment emotion type corresponding to each video content keyword is obtained in the step S6, and the specific processing procedure includes the following steps:
g1, comparing the main comment emotion types corresponding to the comment users corresponding to the video content keywords, and classifying the comment users corresponding to the same main comment emotion types, wherein the same main comment emotion types are recorded as alternative comment emotion types;
g2, counting the number of the alternative comment emotion types corresponding to the video content keywords, wherein if only one alternative comment emotion type corresponds to a certain video content keyword, the comprehensive comment emotion type corresponding to the video content keyword is the alternative comment emotion type;
g3, if the alternative comment emotion types corresponding to the video content keyword are multiple, counting the number of comment users corresponding to each alternative comment emotion type, and further screening out the alternative comment emotion type with the largest number of comment users as the comprehensive comment emotion type corresponding to the video content keyword.
8. The image feature analysis based video clip processing method of claim 1, wherein: in S8, the movie video is edited, and the specific editing method is to use the in-point and out-point of each successfully matched text word in the movie video as the editing start position and the editing end position of each successfully matched text word in the movie video, so as to edit the movie video.
9. An apparatus, characterized by: the system comprises a processor, a memory and a network interface, wherein the memory and the network interface are connected with the processor; the network interface is connected with a nonvolatile memory in the server; the processor, when running, retrieves a computer program from the non-volatile memory via the network interface and runs the computer program via the memory to perform the method of any of claims 1-8.
10. A computer storage medium, characterized in that: the computer storage medium is burned with a computer program, which when run in the memory of the server implements the method of any of the above claims 1-8.
CN202110531926.3A 2021-05-17 2021-05-17 Video clip processing method, apparatus, and computer storage medium based on image feature analysis Withdrawn CN113301383A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110531926.3A CN113301383A (en) 2021-05-17 2021-05-17 Video clip processing method, apparatus, and computer storage medium based on image feature analysis

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110531926.3A CN113301383A (en) 2021-05-17 2021-05-17 Video clip processing method, apparatus, and computer storage medium based on image feature analysis

Publications (1)

Publication Number Publication Date
CN113301383A true CN113301383A (en) 2021-08-24

Family

ID=77322229

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110531926.3A Withdrawn CN113301383A (en) 2021-05-17 2021-05-17 Video clip processing method, apparatus, and computer storage medium based on image feature analysis

Country Status (1)

Country Link
CN (1) CN113301383A (en)

Similar Documents

Publication Publication Date Title
Dhall et al. Emotion recognition in the wild challenge 2013
CN107968959B (en) Knowledge point segmentation method for teaching video
US8676835B2 (en) Annotation system for creating and retrieving media and methods relating to same
CN110442747B (en) Video abstract generation method based on keywords
KR20080058356A (en) Automated rich presentation of a semantic topic
KR20070121810A (en) Synthesis of composite news stories
CN114297439B (en) Short video tag determining method, system, device and storage medium
US10062384B1 (en) Analysis of content written on a board
Dumont et al. Automatic story segmentation for tv news video using multiple modalities
EP2122623B1 (en) Analysing video material
CN112784078A (en) Video automatic editing method based on semantic recognition
CN116361510A (en) Method and device for automatically extracting and retrieving scenario segment video established by utilizing film and television works and scenario
CN113591530A (en) Video detection method and device, electronic equipment and storage medium
CN112468754B (en) Method and device for acquiring pen-recorded data based on audio and video recognition technology
Otani et al. Video summarization using textual descriptions for authoring video blogs
CN116708055B (en) Intelligent multimedia audiovisual image processing method, system and storage medium
JP6603929B1 (en) Movie editing server and program
WO2024030314A1 (en) Search results within segmented communication session content
Nixon et al. Multimodal video annotation for retrieval and discovery of newsworthy video in a news verification scenario
CN113301383A (en) Video clip processing method, apparatus, and computer storage medium based on image feature analysis
CN112468753B (en) Method and device for acquiring and checking record data based on audio and video recognition technology
Haloi et al. Unsupervised story segmentation and indexing of broadcast news video
Liu et al. Major cast detection in video using both audio and visual information
CN114385859A (en) Multi-modal retrieval method for video content
Dong et al. Advanced news video parsing via visual characteristics of anchorperson scenes

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication

Application publication date: 20210824

WW01 Invention patent application withdrawn after publication
DD01 Delivery of document by public notice

Addressee: Wuhan Dexin Yipin e-commerce Co.,Ltd.

Document name: Refund Approval Notice

DD01 Delivery of document by public notice