CN110415274A - A kind of methods of video segmentation - Google Patents
A kind of methods of video segmentation Download PDFInfo
- Publication number
- CN110415274A CN110415274A CN201910689899.5A CN201910689899A CN110415274A CN 110415274 A CN110415274 A CN 110415274A CN 201910689899 A CN201910689899 A CN 201910689899A CN 110415274 A CN110415274 A CN 110415274A
- Authority
- CN
- China
- Prior art keywords
- module
- value
- video
- electrically connected
- picture
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
- G06T7/45—Analysis of texture based on statistical description of texture using co-occurrence matrix computation
-
- 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]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
- H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream, rendering scenes according to MPEG-4 scene graphs
- H04N21/44008—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream, rendering scenes according to MPEG-4 scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics in the video stream
-
- 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
- H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
- H04N21/845—Structuring of content, e.g. decomposing content into time segments
- H04N21/8456—Structuring of content, e.g. decomposing content into time segments by decomposing the content in the time domain, e.g. in time segments
Abstract
The present invention relates to video technique fields, especially a kind of methods of video segmentation, extract texture value in target video image, using texture description video pictures, generate multiple and different characteristic patterns, different characteristic patterns is fused into a comprehensive characteristics figure, comprehensive characteristics figure generation step: the edge locator value size of single feature figure is determined, characteristic pattern with same edge locator value is merged, generate level-one characteristic pattern, determine the centralized positioning value of level-one characteristic pattern, characteristic pattern with identical central locator value is merged, generate secondary characteristics figure, by remaining characteristic pattern according to same characteristic features point Co-factor propagation on secondary characteristics figure, generate comprehensive characteristics figure, according to the characteristic point quantity on comprehensive characteristics figure, video is cut.The present invention also provides a kind of Video segmentation systems.The configuration of the present invention is simple is worthy to be popularized.
Description
Technical field
The present invention relates to video technique field more particularly to a kind of methods of video segmentation.
Background technique
Long video is divided into automatically with the associated short-sighted frequency of event, facilitate user and check and look back video, improve
The service efficiency and experience of user, existing methods of video segmentation are mostly to be manually operated, and divide low efficiency, and segmentation precision is poor.
Summary of the invention
The problem of the purpose of the present invention is to solve above-mentioned technical backgrounds, and a kind of Video segmentation side proposed
Method.
To achieve the goals above, present invention employs following technical solutions:
A kind of methods of video segmentation is designed, is included the following steps:
S1: texture value generates multiple and different features using texture description video pictures in extraction target video image
Figure；
S2: different characteristic patterns is fused into a comprehensive characteristics figure, comprehensive characteristics figure generation step:
(1) the edge locator value size for determining single feature figure, the characteristic pattern with same edge locator value is melted
It closes, generates level-one characteristic pattern；
(2) the centralized positioning value for determining level-one characteristic pattern, the characteristic pattern with identical central locator value is merged, raw
At secondary characteristics figure；
(3) by remaining characteristic pattern according to same characteristic features point Co-factor propagation on secondary characteristics figure, generate comprehensive characteristics figure；
S3: according to the characteristic point quantity on comprehensive characteristics figure, video is cut.
Preferably, in the S1, image texture value extraction step:
(1) video image is imported in picture generator and generates pictorial information；
(2) picture format is changed, so that picture becomes Equalization Histogram, and the grey grade degree of picture is compressed, is generated
The lower picture format of gray level；
(3) element on picture is analyzed using co-occurrence matrix, and element energy is calculated, generate energy from
Non-dramatic song line chart, map discrete curve are the curve graph of texture value thickness value；
(4) the texture value size of particular point in curve graph is extracted as characteristic value.
Preferably, the grey grade degree compression step of picture:
1) picture for needing to carry out gray compression, is chosen；
2) the display default value for, changing picture, reduces the brightness value of picture；
3) the ratio size for, changing picture initial gray value chooses suitable gray value ratio, completes picture ash grade degree pressure
Contracting.
Preferably, in the S3, video cutting step:
(1) it is inserted into comprehensive characteristics figure, characteristic point present in video pictures and the feature on comprehensive characteristics figure in video
Point is compared；
(2) when containing less characteristic point in video pictures, i.e., divisible video generates independent video clip, video
When containing more characteristic point in picture, indivisible video；
(3) video pictures of different time nodes are compared respectively, complete segmentation.
Preferably, the characteristic point is relatively middle sets quantitative value, and quantitative value is sized as feature on comprehensive characteristics figure
The half of point quantity.
The present invention also provides a kind of Video segmentation system, including processor, the processor is electrically connected with data and connects
Module is received, the data reception module is electrically connected with picture converter, and the processor is electrically connected with memory module, described
Processor is electrically connected with data processing unit, and the data processing unit includes microprocessor, the microprocessor electricity
Property be connected with memory, the microprocessor is electrically connected with texture blending module, and the texture blending module is electrically connected
There is picture recognition module, the picture recognition module is electrically connected with signal conversion module, and the microprocessor is electrically connected
There is synthesis unit, the microprocessor is electrically connected with ratio module, and data reception module transmits received pictorial information
Into processor, after storing, pictorial information is transmitted in picture recognition module for processor, and picture recognition module is to transmission
Pictorial information is identified, extracts the data texturing value in pictorial information, the input of data texturing value by texture blending module
In synthesis unit, new characteristic pattern and the characteristic pattern of acquisition are carried out characteristic point quantity by the new characteristic pattern of generation, ratio module
Compare.
Preferably, the synthesis unit includes location information acquisition module, and the location information acquisition module is electrically connected
There is locating module, the locating module is electrically connected with matching module, and the matching module, which is electrically connected with, transfers module, described
It transfers module and is electrically connected with merging module, location information acquisition module acquisition characteristics point generates the location information of feature image,
Locating module positions the feature image of generation, is moved the feature image of location information having the same by transferring module
It moves to identical position, comprehensive characteristics figure is synthesized by merging module.
Preferably, the ratio module includes microchip, and the microchip is electrically connected with database, described miniature
Chip is electrically connected with characteristic point acquisition module, and the microchip is electrically connected with counting module, and the microchip is electrical
It is connected with definite value module, the microchip is electrically connected with determination module, and the acquisition of characteristic point acquisition module inputs on picture
Characteristic point quantity, counting module count the quantity of acquisition, and determination module will be in obtained population size and definite value module
The definite value size of setting compares, and determines whether to be split.
A kind of methods of video segmentation proposed by the present invention, beneficial effect are:
1, it is counted by quantity of the counting module to acquisition, determination module is by obtained population size and definite value module
The definite value size of interior setting compares, and when the quantity of characteristic point is lower than setting value, is split, otherwise does not divide, and divides
Efficiency is fast, and segmentation is accurate.
Detailed description of the invention
Fig. 1 is a kind of system block diagram of methods of video segmentation proposed by the present invention；
Fig. 2 is the system block diagram of synthesis unit in the present invention；
Fig. 3 is the system block diagram of ratio module in the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.
Referring to Fig.1-3, a kind of methods of video segmentation, includes the following steps:
S1: texture value generates multiple and different features using texture description video pictures in extraction target video image
Figure, image texture value extraction step:
(1) video image is imported in picture generator and generates pictorial information；
(2) picture format is changed, so that picture becomes Equalization Histogram, and the grey grade degree of picture is compressed, is generated
Gray level lower picture format is handled picture after gray compression is convenient, the grey grade degree compression step of picture:
1) picture for needing to carry out gray compression, is chosen；
2) the display default value for, changing picture, reduces the brightness value of picture；
3) the ratio size for, changing picture initial gray value chooses suitable gray value ratio, completes picture ash grade degree pressure
Contracting；
(3) element on picture is analyzed using co-occurrence matrix, and element energy is calculated, generate energy from
Non-dramatic song line chart, map discrete curve are the curve graph of texture value thickness value；
(4) the texture value size of particular point in curve graph is extracted as characteristic value；
S2: different characteristic patterns is fused into a comprehensive characteristics figure, comprehensive characteristics figure generation step:
(1) the edge locator value size for determining single feature figure, the characteristic pattern with same edge locator value is melted
It closes, generates level-one characteristic pattern；
(2) the centralized positioning value for determining level-one characteristic pattern, the characteristic pattern with identical central locator value is merged, raw
At secondary characteristics figure；
(3) by remaining characteristic pattern according to same characteristic features point Co-factor propagation on secondary characteristics figure, generate comprehensive characteristics figure；
S3: according to the characteristic point quantity on comprehensive characteristics figure, video is cut, video cutting step:
(1) it is inserted into comprehensive characteristics figure, characteristic point present in video pictures and the feature on comprehensive characteristics figure in video
Point is compared；
(2) when containing less characteristic point in video pictures, i.e., divisible video generates independent video clip, video
When containing more characteristic point in picture, indivisible video；
(3) video pictures of different time nodes are compared respectively, complete segmentation, characteristic point is relatively middle to set quantity
Value, quantitative value are sized as the half of characteristic point quantity on comprehensive characteristics figure.
The present invention also provides a kind of Video segmentation systems, including processor, processor to be electrically connected with data reception
Block, data reception module are the pictorial informations for receiving the transmission of picture converter, and data reception module is electrically connected with picture conversion
Device, picture converter are to convert video pictures in picture, and processor is electrically connected with memory module, and processor is electrically connected with
Data processing unit, data processing unit are handled the pictorial information of transmission, and data processing unit includes miniature processing
Device, microprocessor are electrically connected with memory.
Microprocessor is electrically connected with texture blending module, and texture blending module is carried out to the texture information on picture
It extracts, texture blending module is electrically connected with picture recognition module, and picture recognition module is identification picture, picture recognition module electricity
Property be connected with signal conversion module, microprocessor is electrically connected with synthesis unit, and synthesis unit is that the characteristic pattern generated carries out
Synthesis, microprocessor are electrically connected with ratio module, and received pictorial information is transmitted in processor by data reception module,
After storing, pictorial information is transmitted in picture recognition module for processor, picture recognition module to the pictorial information of transmission into
Row identification extracts the data texturing value in pictorial information by texture blending module, and data texturing value inputs in synthesis unit,
New characteristic pattern and the characteristic pattern of acquisition progress characteristic point quantity are compared and work as characteristic point by the new characteristic pattern generated, ratio module
Quantity when being lower than setting value, be split, otherwise do not divide.
Synthesis unit includes location information acquisition module, and location information acquisition module is the position letter that acquisition generates characteristic pattern
Breath, location information acquisition module are electrically connected with locating module, and the effect of locating module is fixed character figure, and locating module is electrical
It is connected with matching module, matching module is to be attached the characteristic pattern of same position, and matching module, which is electrically connected with, transfers mould
Block transfers module and is electrically connected with merging module, and merging module is to merge all characteristic patterns, generates comprehensive characteristics figure, position
The location information that information acquisition module acquisition characteristics point generates feature image is set, locating module determines the feature image of generation
Position, is moved to identical position for the feature image of location information having the same by transferring module, is closed by merging module
At comprehensive characteristics figure.
Ratio module includes microchip, and microchip is electrically connected with database, and microchip is electrically connected with feature
Point acquisition module, microchip are electrically connected with counting module, and counting module is the quantity for calculating characteristic point on picture, minicore
Piece is electrically connected with definite value module, and definite value module is setting reduced value, and microchip is electrically connected with determination module, and characteristic point is adopted
Collect the characteristic point quantity on module acquisition input picture, counting module counts the quantity of acquisition, and determination module will obtain
Population size and definite value module in set definite value size compare, when the quantity of characteristic point be lower than setting value when, carry out
Segmentation, on the contrary do not divide.
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, according to the technique and scheme of the present invention and its
Inventive concept is subject to equivalent substitution or change, should be covered by the protection scope of the present invention.
Claims (8)
1. a kind of methods of video segmentation, which comprises the steps of:
S1: texture value generates multiple and different characteristic patterns using texture description video pictures in extraction target video image；
S2: different characteristic patterns is fused into a comprehensive characteristics figure, comprehensive characteristics figure generation step:
(1) the edge locator value size for determining single feature figure, the characteristic pattern with same edge locator value is merged, raw
At level-one characteristic pattern；
(2) the centralized positioning value for determining level-one characteristic pattern, the characteristic pattern with identical central locator value is merged, and generates two
Grade characteristic pattern；
(3) by remaining characteristic pattern according to same characteristic features point Co-factor propagation on secondary characteristics figure, generate comprehensive characteristics figure；
S3: according to the characteristic point quantity on comprehensive characteristics figure, video is cut.
2. a kind of methods of video segmentation according to claim 1, which is characterized in that in the S1, image texture value is extracted
Step:
(1) video image is imported in picture generator and generates pictorial information；
(2) picture format is changed, so that picture becomes Equalization Histogram, and the grey grade degree of picture is compressed, gray scale is generated
The lower picture format of grade；
(3) element on picture is analyzed using co-occurrence matrix, and element energy is calculated, it is bent to generate spread in energy
Line chart, map discrete curve are the curve graph of texture value thickness value；
(4) the texture value size of particular point in curve graph is extracted as characteristic value.
3. a kind of methods of video segmentation according to claim 2, which is characterized in that the grey grade degree of the picture compresses step
It is rapid:
1) picture for needing to carry out gray compression, is chosen；
2) the display default value for, changing picture, reduces the brightness value of picture；
3) the ratio size for, changing picture initial gray value chooses suitable gray value ratio, completes the compression of picture ash grade degree.
4. a kind of methods of video segmentation according to claim 1, which is characterized in that in the S3, video cutting step:
(1) it is inserted into comprehensive characteristics figure in video, characteristic point present in video pictures and the feature on comprehensive characteristics figure click through
Row compares；
(2) when containing less characteristic point in video pictures, i.e., divisible video generates independent video clip, video pictures
In contain more characteristic point when, indivisible video；
(3) video pictures of different time nodes are compared respectively, complete segmentation.
5. a kind of methods of video segmentation according to claim 4, which is characterized in that the characteristic point is relatively middle to set quantity
Value, quantitative value are sized as the half of characteristic point quantity on comprehensive characteristics figure.
6. a kind of Video segmentation system described in one of -5 according to claim 1, which is characterized in that including processor, the place
Reason device is electrically connected with data reception module, and the data reception module is electrically connected with picture converter, the processor electricity
Property is connected with memory module, and the processor is electrically connected with data processing unit, and the data processing unit includes miniature place
Device is managed, the microprocessor is electrically connected with memory, and the microprocessor is electrically connected with texture blending module, described
Texture blending module is electrically connected with picture recognition module, and the picture recognition module is electrically connected with signal conversion module, institute
It states microprocessor and is electrically connected with synthesis unit, the microprocessor is electrically connected with ratio module, data reception module
Received pictorial information is transmitted in processor, after storing, pictorial information is transmitted in picture recognition module for processor,
Picture recognition module identifies the pictorial information of transmission, and the texture number in pictorial information is extracted by texture blending module
According to value, data texturing value is inputted in synthesis unit, the new characteristic pattern of generation, and ratio module is by the spy of new characteristic pattern and acquisition
Sign figure carries out characteristic point quantity and compares.
7. a kind of Video segmentation system according to claim 6, which is characterized in that the synthesis unit includes location information
Acquisition module, the location information acquisition module are electrically connected with locating module, and the locating module is electrically connected with matching mould
Block, the matching module, which is electrically connected with, transfers module, and the module of transferring is electrically connected with merging module, location information acquisition
Module acquisition characteristics point generates the location information of feature image, and locating module positions the feature image of generation, passes through tune
The feature image of location information having the same is moved to identical position by modulus block, synthesizes comprehensive characteristics by merging module
Figure.
8. a kind of Video segmentation system according to claim 6, which is characterized in that the ratio module includes minicore
Piece, the microchip are electrically connected with database, and the microchip is electrically connected with characteristic point acquisition module, described miniature
Chip is electrically connected with counting module, and the microchip is electrically connected with definite value module, and the microchip is electrically connected with
Determination module, the acquisition of characteristic point acquisition module input the characteristic point quantity on picture, and counting module counts the quantity of acquisition
Number, determination module compares the definite value size set in obtained population size and definite value module, determines whether point
It cuts.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910689899.5A CN110415274B (en) | 2019-07-29 | 2019-07-29 | Video segmentation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910689899.5A CN110415274B (en) | 2019-07-29 | 2019-07-29 | Video segmentation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110415274A true CN110415274A (en) | 2019-11-05 |
| CN110415274B CN110415274B (en) | 2023-09-15 |
Family
ID=68363821
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910689899.5A Active CN110415274B (en) | 2019-07-29 | 2019-07-29 | Video segmentation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110415274B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107437076A (en) * | 2017-08-02 | 2017-12-05 | 陈雷 | The method and system that scape based on video analysis does not divide |
| US20190114774A1 (en) * | 2017-10-16 | 2019-04-18 | Adobe Systems Incorporated | Generating Image Segmentation Data Using a Multi-Branch Neural Network |
-
2019
- 2019-07-29 CN CN201910689899.5A patent/CN110415274B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107437076A (en) * | 2017-08-02 | 2017-12-05 | 陈雷 | The method and system that scape based on video analysis does not divide |
| US20190114774A1 (en) * | 2017-10-16 | 2019-04-18 | Adobe Systems Incorporated | Generating Image Segmentation Data Using a Multi-Branch Neural Network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110415274B (en) | 2023-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102077580B (en) | Display control device, display control method | |
| CN111754302B (en) | Video live broadcast interface commodity display intelligent management system based on big data | |
| EP1221662A3 (en) | Image retrieval method | |
| CN109871762B (en) | Face recognition model evaluation method and device | |
| CN103530649A (en) | Visual searching method applicable mobile terminal | |
| CN102655607B (en) | For adding up the method and apparatus of the audience ratings of target audience | |
| CN111291825A (en) | Focus classification model training method and device, computer equipment and storage medium | |
| CN113221732B (en) | Realization method for precisely manufacturing intelligent certificate card by big data based on face recognition | |
| CN110659817A (en) | Data processing method and device, machine readable medium and equipment | |
| CN112307232A (en) | Intelligent classification storage processing method for big data content | |
| CN116562270A (en) | Natural language processing system supporting multi-mode input and method thereof | |
| CN105740777B (en) | Information processing method and device | |
| CN113192028B (en) | Quality evaluation method and device for face image, electronic equipment and storage medium | |
| CN110415274A (en) | A kind of methods of video segmentation | |
| CN101587512A (en) | Appraising method for gap filling questions in computer auxiliary test system | |
| US20020023952A1 (en) | Card exchanging device, card exchanging method, and recording media | |
| CN112287945A (en) | Screen fragmentation determination method and device, computer equipment and computer readable storage medium | |
| CN110414593A (en) | Image processing method and device, processor, electronic equipment and storage medium | |
| CN109035419A (en) | A kind of social contact method and system based on AR technology | |
| CN111209860B (en) | Video attendance system and method based on deep learning and reinforcement learning | |
| CN112365248B (en) | Method and system for analyzing digital currency transaction path | |
| CN113269283A (en) | Self-service card sender of all-purpose card identification card | |
| CN111507588A (en) | Intelligent recommendation system for intellectual property service | |
| CN101587514A (en) | Grading system for cloze test in computer aided examination system | |
| CN111783592B (en) | Face recognition method combining incomplete face information with position data |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |