CN102957930B - Method and system for automatically identifying 3D (Three-Dimensional) format of digital content - Google Patents
Method and system for automatically identifying 3D (Three-Dimensional) format of digital content Download PDFInfo
- Publication number
- CN102957930B CN102957930B CN201210319607.7A CN201210319607A CN102957930B CN 102957930 B CN102957930 B CN 102957930B CN 201210319607 A CN201210319607 A CN 201210319607A CN 102957930 B CN102957930 B CN 102957930B
- Authority
- CN
- China
- Prior art keywords
- video information
- continuity
- met
- judge
- middle column
- 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.)
- Active
Links
Landscapes
- Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
Abstract
The invention discloses a method and a system for automatically identifying 3D (Three-Dimensional) format of digital content. The method comprises the following steps of: after receiving video information, judging the 3D format or the 2D format of the video information according to the continuity of middle rows of the video information and/or middle lines and/or virtual middle rows of the video information; and carrying out signal reorganization processing on the video information according to the 3D format or the 2D format of the video information so as to obtain recovered video frames. By utilizing the method and the system, the automatic judgment on the 3D format of digital content is realized, and the reproduction of the frames is further realized, so that the efficiency in 3D play is improved, and multiple problems in manual setup are avoided.
Description
Technical field
The present invention relates to 3D video technique field, refer to a kind of digital content 3D form automatic identifying method and system especially.
Background technology
Along with the development of 3D industry, emerge a large amount of 3D video formats, be mainly divided into two large classes, comprise: frame compatible format and non-frame compatible format.Wherein, frame compatible format adopts widely due to the flow process and obtaining of the current 2D transfer of data of compatibility, reception and process.The extraction different according to right and left eyes two-way image and joining method are distinguished, frame compatible format can be divided into: left and right (Side by Side, SBS) form, up and down (Top-and-Bottom, TaB) form, SENSIO high-fidelity 3D (SENSIO HiFi3D) form, gridiron pattern (Checkerboard) form and 3D floor (3D Tile) form.Play System needs to carry out effectively distinguishing to determine subsequent treatment to above-mentioned five kinds of forms and 2D form in actual applications.
Add respective identification (Flag) in the relevant criterion such as current dynamic image expert group 2 (Moving Pictures Experts Group2, MPEG2) and MPEG4 to distinguish SBS form, TaB form and SENSIO HiFi3D form.According to Flag wherein, client end of playing back, after carrying out corresponding decoding to video flowing, can determine that subsequent treatment is correctly to recover or to show 3D content.Current audio/video encoding standard (Audio Video coding Standard, AVS standard) SBS, TaB form can be supported, further actively pushing forward to make AVS standard also can support SENSIO HiFi3D form, Checkerboard form and 3D Tile form.In addition, part Play System such as: the playout software of Microsoft, and the LEC3DS Play System of LEONIS, also support user oneself is according to the 3D information standard specification of respective company, adds format information and set up corresponding format database in this locality for material.
But there is limitation in the application in said method: after video data decoding, format information is difficult to be transmitted to subordinate by transmission channel on the one hand, and its main cause does not have current standard and resource, and cost is relatively high; Rely on the other hand the also very inconvenient and as easy as rolling off a log generation misoperation of mode that user adds voluntarily.Once database lost will be difficult to perform subsequent operation.For real-time 3D format conversion system, after pattern of the input converts, manual setting is easy to produce mistake and because its retardance is for the process of material or viewing effect all can produce negative impact.
Summary of the invention
In view of this, main purpose of the present invention is to provide a kind of digital content 3D form automatic identifying method and system, can solve the problem that digital content 3D form cannot automatically judge and process.
For achieving the above object, technical scheme of the present invention is achieved in that
The invention discloses a kind of digital content 3D form automatic identifying method, described method comprises:
After receiving video information, detect the continuity of the middle column of video information, then judge that video information is as left and right (SBS) form or SENSIO HiFi3D form as met noncontinuity, the continuity of the middle row of video information is then detected as met continuity, then judge that video information is as upper and lower (TaB) form as met noncontinuity, the virtual middle column of video information is then constructed as met continuity, and detect its continuity, then judge that video information is as 2D form as met continuity, otherwise judge that video information is as gridiron pattern (Checkerboard) form;
3D form belonging to video information or 2D form, carry out signal restructuring process to video information, obtain the video pictures reduced.
Wherein, described judge that video information is as SBS form or SENSIO HiFi3D form after, described method also comprises: distinguish SBS form or SENSIO HiFi3D form by quincunx reverse sawtooth algorithm (Quincux).
Wherein, the virtual middle column of described structure video information, is specially the mode constructing virtual middle column according to De-CheckBoard.
Wherein, after the 3D form judged belonging to video information or 2D form, described method also comprises: utilize the mode that similitude confirms, checks the 3D form belonging to video information or 2D form.
The invention also discloses a kind of digital content 3D form automatic recognition system, described system comprises: signal receiving module, form determination module and signal processing module, wherein,
Described form determination module, after receiving video information that signal receiving module sends, detect the continuity of the middle column of video information, then judge that video information is as SBS form or SENSIO HiFi3D form as met noncontinuity, the continuity of the middle row of video information is then detected as met continuity, then judge that video information is as TaB form as met noncontinuity, the virtual middle column of video information is then constructed as met continuity, and detect its continuity, then judge that video information is as 2D form as met continuity, otherwise judge that video information is as Checkerboard form, 3D form belonging to video information or 2D form and video information are sent to signal processing module,
Described signal processing module, for the 3D form belonging to video information or 2D form, carries out signal restructuring process to video information, obtains the video pictures reduced.
Wherein, described form determination module, also for after judging that video information is as SBS form or SENSIOHiFi3D form, distinguishes SBS form or SENSIO HiFi3D form by Quincux.
Wherein, the virtual middle column of described form determination module structure video information, is specially the mode constructing virtual middle column according to De-CheckBoard.
Wherein, described form determination module, also for after the 3D form judged belonging to video information or 2D form, utilizes the mode that similitude confirms, checks the 3D form belonging to video information or 2D form.
Digital content 3D form automatic identifying method provided by the present invention and system, after receiving video information, according to the middle column of video information and/or the continuity of middle row and/or virtual middle column, judge the 3D form belonging to video information or 2D form; 3D form belonging to video information or 2D form, carry out signal restructuring process to video information, obtain the video pictures reduced.By said method and system, the automatic judgement of digital content 3D form can be realized, and realize the reproduction of picture further, improve the efficiency that 3D plays, avoid the problems that existence is manually set.
Accompanying drawing explanation
Fig. 1 is a kind of digital content 3D form of the present invention automatic identifying method schematic flow sheet;
Fig. 2 is the schematic diagram that the continuity of carrying out middle column for SBS form or SENSIO HiFi3D form detects;
Fig. 3 is the schematic diagram that the continuity of carrying out middle row for TaB form detects;
Fig. 4 is the schematic diagram that the continuity of carrying out virtual middle column for Checkerboard form detects;
Fig. 5 is the principle of compositionality schematic diagram of 3D Tile form;
Fig. 6 is the schematic diagram of helical scan type search;
Fig. 7 is the schematic diagram of three cloth search methods;
Fig. 8 is a kind of digital content 3D form of the present invention automatic recognition system.
Embodiment
Basic thought of the present invention is: after receiving video information, according to the middle column of video information and/or the continuity of middle row and/or virtual middle column, judges the 3D form belonging to video information or 2D form; 3D form belonging to video information or 2D form, carry out signal restructuring process to video information, obtain the video pictures reduced.
Below in conjunction with the drawings and specific embodiments, the technical solution of the present invention is further elaborated.
Fig. 1 is a kind of digital content 3D form of the present invention automatic identifying method schematic flow sheet, and as shown in Figure 1, described method comprises:
Step 101, after receiving video information, according to the middle column of video information and/or the continuity of middle row and/or virtual middle column, judges the 3D form belonging to video information or 2D form;
Step 102, the 3D form belonging to video information or 2D form, carry out signal restructuring process to video information, obtain the video pictures reduced.
Further, after step 101, described method also comprises: utilize similitude to confirm the mode of (shiftestimate), check the 3D form belonging to video information or 2D form.
Concrete, in described step 101, according to the middle column of video information and/or the continuity of middle row and/or virtual middle column, judge the 3D form belonging to video information or 2D form, specifically comprise the following steps:
Step 101a, detects the continuity of the middle column of video information, then performs step 101b, otherwise judge that video information is as SBS form or SENSIO HiFi3D form as met continuity;
Concrete, Fig. 2 is the schematic diagram that the continuity of carrying out middle column for SBS form or SENSIO HiFi3D form detects, and as shown in Figure 2, if meet noncontinuity, can judge that video information is as SBS form or SENSIO HiFi3D form.
Further, after judging that video information is as SBS form or SENSIO HiFi3D form, described method also comprises: distinguish SBS form or SENSIO HiFi3D form by quincunx reverse sawtooth algorithm (Quincux).
Step 101b, detects the continuity of the middle row of video information, then performs step 101c, otherwise judge that video information is as TaB form as met continuity;
Concrete, Fig. 3 is the schematic diagram that the continuity of carrying out middle row for TaB form detects, and as shown in Figure 3, if meet noncontinuity, can judge that video information is as TaB form.
Step 101c, the virtual middle column of structure video information, and detect its continuity, then judge that video information is as 2D form as met continuity, otherwise judge that video information is as Checkerboard form.
Concrete, the virtual middle column of described structure video information, is specially the mode constructing virtual middle column according to De-CheckBoard.So-called virtual middle column is reorganized data, and structure can detect successional middle column.Fig. 4 is the schematic diagram that the continuity of carrying out virtual middle column for Checkerboard form detects, and as shown in Figure 4, if meet noncontinuity, can judge that video information is as Checkerboard form.In said process, described otherwise specifically refer to meet noncontinuity.
Further, Fig. 5 is the principle of compositionality schematic diagram of 3D Tile form, as shown in Figure 5, as long as detect the continuity arranged between L and R1 and the continuity of going between T and (R2, R3).
In addition, in the present invention, successional detection is specially: edge detection algorithm.It is roughly divided into a few class, the differential method, Surface Fitting and optimum operator method, multiple dimensioned algorithm etc.All needed to adopt rim detection to carry out 3D format identification by the known the present invention of Fig. 2 to Fig. 5.Adopt derivative to describe the change of continuous function in infinitesimal calculus, the change of image function represents by the gradient pointing to function maximum growth direction.Image f (x, y) is defined as in the gradient at position (x, y) place
The amplitude of gradient
Deflection
gradient direction is vertical with edge direction.Digital picture is discrete in essence, and therefore will obtain partial derivative can be similar to by difference.Common rim detection giving gradient has Robert, Prewitt, Sobel, Kirsch operator etc., and their general amount of calculation is little, simple to operate.Wherein Sobel operator is applied to horizontal and vertical edge usually.Meet the demand that our image display detects.
Sobel edge detection algorithm obtains horizontal gradient and vertical gradient in mainly being inputted by image, then carries out gradient combination, carries out the image that threshold processing then can obtain output afterwards according to thresholding T.Wherein, Sobel operator is mainly utilized to make rim detection.Its essence is a discrete difference operator, be used for the gray approximation of computed image luminance function.Any point in the picture uses, and all will produce corresponding gray scale vector.Generally comprise the matrix of following two groups of 3X3, represent laterally (table 1) and longitudinally (table 2) operator respectively, itself and image are made planar convolution, horizontal and vertical brightness difference approximation can be obtained respectively.
The horizontal and vertical gray value of each pixel of image is combined by gradient formula, can calculate the gray scale size of this point:
general in order to improve computational efficiency, we adopt more approximate formula: | G|=|G
x|+| G
y|.And then reach this phenomenon Edge detected of extreme value by, left and right adjoint point intensity-weighted difference upper and lower according to pixel in edge.
Similitude described in the present invention confirms to be specially to search for completely from center, in order to improve search speed, can adopt fast search algorithm.
Search for each its SAE of calculating (absolute error and) value in region of search (to current macro+s) completely, advantage one finds SAE minimum in region of search surely, and shortcoming is that amount of calculation is large, needs (2s+1)
2secondary computation measure.The complete search strategy of usual employing has two kinds: one is grating scanning type search, and from the most upper left corner of region of search, raster scan also calculates all positions, and amount of calculation is larger.And for our actual 3D format identification application, major part motion vector is all around central point, therefore the search of another kind of helical scan type can be reduced to further, from center, the form of block of spiral is adopted to search for clockwise, calculating more rearward more may be interrupted, thus saves and assess the cost.Fig. 6 is the schematic diagram of helical scan type search.
And due to computational resource in practical application or power limited, adopting in the process of helical scan type region of search in employing, employing fast algorithm that can be suitable, only calculates the SAE of part point at intra-zone, thus reduction assesses the cost greatly.General employing TTS carries out fast search, i.e. three step search algorithm.Fig. 7 is the schematic diagram of three cloth search methods, as shown in Figure 7, first choose around central point 8 sampling points (usually get ± 2
n-18 sampling points) calculate its SAE, and reference number 1.In 8 points, select the point that SAE is minimum, reduce
detection range, to continue around this point of search 8 sampling points calculate its SAE, and reference number 2.Distance repeats this operation further, until cannot reduce further.Thus amount of calculation significantly reduces.
In order to calculate the size of side-play amount, need to calculate its energy.The tolerance of usual employing energy has three kinds of form: MSE (Mean Square Error), MAE (mean absolute error) and SAE (absolute error and).Different tolerance affects the accuracy of computation complexity and estimation.Wherein, SAE applies the most widely because it has relatively low computation complexity.Formula is:
what adopt in spiral complete fast search is SAE.
Fig. 8 is the structural representation of a kind of digital content 3D form of the present invention automatic recognition system, and as shown in Figure 8, described system comprises: signal receiving module 81, form determination module 82 and signal processing module 83, wherein,
Described form determination module 82, after receiving video information that signal receiving module 81 sends, according to the middle column of video information and/or the continuity of middle row and/or virtual middle column, judge 3D form belonging to video information or 2D form, the 3D form belonging to video information or 2D form and video information are sent to signal processing module 83;
Described signal processing module 83, for the 3D form belonging to video information or 2D form, carries out signal restructuring process to video information, obtains the video pictures reduced.
Concrete, described form determination module 82 is according to the middle column of video information and/or centre is capable and/or the continuity of virtual middle column, judges the 3D form belonging to video information or 2D form, specifically comprises:
Form determination module 82 detects the continuity of the middle column of video information, then judge that video information is as SBS form or SENSIO HiFi3D form as met noncontinuity, the continuity of the middle row of video information is then detected as met continuity, then judge that video information is as TaB form as met noncontinuity, the virtual middle column of video information is then constructed as met continuity, and detect its continuity, then judge that video information is as 2D form as met continuity, otherwise judge that video information is as Checkerboard form.
Wherein, the virtual middle column of described form determination module structure video information, is specially the mode constructing virtual middle column according to De-CheckBoard.
Further, described form determination module 82, also for after judging that video information is as SBS form or SENSIO HiFi3D form, distinguishes SBS form or SENSIO HiFi3D form by Quincux.
Further, described form determination module 82, also for after the 3D form judged belonging to video information or 2D form, utilizes the mode that similitude confirms, checks the 3D form belonging to video information or 2D form.
The above, be only preferred embodiment of the present invention, be not intended to limit protection scope of the present invention.
Claims (6)
1. a digital content 3D form automatic identifying method, is characterized in that, described method comprises:
After receiving video information, detect the continuity of the middle column of video information, then judge that video information is as left and right (SBS) form or SENSIO high-fidelity 3D (SENSIO HiFi 3D) form as met noncontinuity, SBS form or SENSIO HiFi 3D form is distinguished by quincunx reverse sawtooth algorithm (Quincux), the continuity of the middle row of video information is then detected as met continuity, then judge that video information is as upper and lower (TaB) form as met noncontinuity, the virtual middle column of video information is then constructed as met continuity, and detect its continuity, then judge that video information is as 2D form as met continuity, otherwise judge that video information is as gridiron pattern (Checkerboard) form,
3D form belonging to video information or 2D form, carry out signal restructuring process to video information, obtain the video pictures reduced.
2. method according to claim 1, is characterized in that, the virtual middle column of described structure video information, is specially the mode constructing virtual middle column according to De-CheckBoard.
3. method according to claim 1, is characterized in that, after the 3D form judged belonging to video information or 2D form, described method also comprises: utilize the mode that similitude confirms, checks the 3D form belonging to video information or 2D form.
4. a digital content 3D form automatic recognition system, is characterized in that, described system comprises: signal receiving module, form determination module and signal processing module, wherein,
Described form determination module, after receiving video information that signal receiving module sends, detect the continuity of the middle column of video information, then judge that video information is as SBS form or SENSIO HiFi 3D form as met noncontinuity, the continuity of the middle row of video information is then detected as met continuity, then judge that video information is as TaB form as met noncontinuity, the virtual middle column of video information is then constructed as met continuity, and detect its continuity, then judge that video information is as 2D form as met continuity, otherwise judge that video information is as Checkerboard form, 3D form belonging to video information or 2D form and video information are sent to signal processing module,
Described form determination module, also for after judging that video information is as SBS form or SENSIO HiFi 3D form, distinguishes SBS form or SENSIO HiFi 3D form by Quincux;
Described signal processing module, for the 3D form belonging to video information or 2D form, carries out signal restructuring process to video information, obtains the video pictures reduced.
5. system according to claim 4, is characterized in that, the virtual middle column of described form determination module structure video information, is specially the mode constructing virtual middle column according to De-CheckBoard.
6. system according to claim 4, it is characterized in that, described form determination module, also for after the 3D form judged belonging to video information or 2D form, utilize the mode that similitude confirms, the 3D form belonging to video information or 2D form are checked.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210319607.7A CN102957930B (en) | 2012-09-03 | 2012-09-03 | Method and system for automatically identifying 3D (Three-Dimensional) format of digital content |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210319607.7A CN102957930B (en) | 2012-09-03 | 2012-09-03 | Method and system for automatically identifying 3D (Three-Dimensional) format of digital content |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102957930A CN102957930A (en) | 2013-03-06 |
| CN102957930B true CN102957930B (en) | 2015-03-11 |
Family
ID=47766080
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210319607.7A Active CN102957930B (en) | 2012-09-03 | 2012-09-03 | Method and system for automatically identifying 3D (Three-Dimensional) format of digital content |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102957930B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105898270A (en) * | 2015-12-27 | 2016-08-24 | 乐视致新电子科技（天津）有限公司 | Video format distinguishing method and system |
| CN105898269A (en) * | 2015-12-27 | 2016-08-24 | 乐视致新电子科技（天津）有限公司 | Video play method and device |
| CN106131528B (en) * | 2016-06-23 | 2018-07-10 | 福建天泉教育科技有限公司 | The recognition methods of 3D video formats and system |
| CN108830198A (en) * | 2018-05-31 | 2018-11-16 | 上海玮舟微电子科技有限公司 | Recognition methods, device, equipment and the storage medium of video format |
| CN109672881A (en) * | 2019-01-04 | 2019-04-23 | 南京大学 | A kind of method of automatic identification 2D/3D video format |
| CN112004162B (en) * | 2020-09-08 | 2022-06-21 | 宁波视睿迪光电有限公司 | Online 3D content playing system and method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102231829A (en) * | 2011-06-27 | 2011-11-02 | 深圳超多维光电子有限公司 | Display format identification method and device of video file as well as video player |
| CN102395037A (en) * | 2011-06-30 | 2012-03-28 | 深圳超多维光电子有限公司 | Format recognition method and device |
| CN102461185A (en) * | 2009-06-23 | 2012-05-16 | 三星电子株式会社 | Method and apparatus for automatic transformation of three-dimensional video |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110164110A1 (en) * | 2010-01-03 | 2011-07-07 | Sensio Technologies Inc. | Method and system for detecting compressed stereoscopic frames in a digital video signal |
-
2012
- 2012-09-03 CN CN201210319607.7A patent/CN102957930B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102461185A (en) * | 2009-06-23 | 2012-05-16 | 三星电子株式会社 | Method and apparatus for automatic transformation of three-dimensional video |
| CN102231829A (en) * | 2011-06-27 | 2011-11-02 | 深圳超多维光电子有限公司 | Display format identification method and device of video file as well as video player |
| CN102395037A (en) * | 2011-06-30 | 2012-03-28 | 深圳超多维光电子有限公司 | Format recognition method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102957930A (en) | 2013-03-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102957930B (en) | Method and system for automatically identifying 3D (Three-Dimensional) format of digital content | |
| CN101282492B (en) | Method for regulating display depth of three-dimensional image | |
| CN101980545B (en) | Method for automatically detecting 3DTV video program format | |
| CN103810685B (en) | A kind of super-resolution processing method of depth map | |
| JP5575891B2 (en) | Automatic 3D video format conversion method and apparatus | |
| CN102006425B (en) | Method for splicing video in real time based on multiple cameras | |
| CN102547344B (en) | Video format identification method and video format identification device | |
| EP2560398A2 (en) | Method and apparatus for correcting errors in stereo images | |
| CN101969564B (en) | Upsampling method for depth video compression of three-dimensional television | |
| CN1226786A (en) | View offset estimation for stereo scopic video coding | |
| WO2001004837A3 (en) | Method and apparatus for detecting independent motion in three-dimensional scenes | |
| CN101600061A (en) | De-interlaced method of video motion-adaptive and device | |
| CN105210374B (en) | The control of super-resolution is improved to frequency using characteristics of image | |
| CN102547350B (en) | Method for synthesizing virtual viewpoints based on gradient optical flow algorithm and three-dimensional display device | |
| CN101840574A (en) | Depth estimation method based on edge pixel features | |
| TW201143358A (en) | Method for distinguishing a 3D image from a 2D image and for identifying the presence of a 3D image format by feature correspondence determination | |
| CN105069804A (en) | Three-dimensional model scanning reconstruction method based on smartphone | |
| US8718331B2 (en) | Image detecting apparatus and method thereof | |
| CN101557534B (en) | Method for generating disparity map from video close frames | |
| CN103440664A (en) | Method, system and computing device for generating high-resolution depth map | |
| CN102447917A (en) | Three-dimensional image matching method and equipment thereof | |
| KR20110113923A (en) | Image converting device and three dimensional image display device including the same | |
| CN103077542A (en) | Compression method for interest region of depth map | |
| CN102883170B (en) | Image processing apparatus and image processing method | |
| CN104980726A (en) | Binocular video stereo matching method based on motion vector |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |