CN1327701C - Method and relative system for high efficiency ad detection in video signal - Google Patents
Method and relative system for high efficiency ad detection in video signal Download PDFInfo
- Publication number
- CN1327701C CN1327701C CNB2004100558090A CN200410055809A CN1327701C CN 1327701 C CN1327701 C CN 1327701C CN B2004100558090 A CNB2004100558090 A CN B2004100558090A CN 200410055809 A CN200410055809 A CN 200410055809A CN 1327701 C CN1327701 C CN 1327701C
- Authority
- CN
- China
- Prior art keywords
- frame
- advertising segment
- search
- vision signal
- reference frame
- 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
Images
Abstract
The present invention provides a method and a relevant system capable of detecting advertisements in a video signal. An advertisement segment can be inserted in a normal program in video signals provided by broadcasting and television media, so that pictures with consecutive frames at the advertisement insertion point are not continuous; after the advertisement segment is finished, part of the segment of the previous section of the normal program is replayed. The present invention has the technical scheme that whether the pictures with two consecutive frames in the video signals are continuous is firstly detected, and whether the frame of the similar picture exists is searched backwards from the discontinuous frame of the picture. If the frame of the similar picture exists, then a frame of the advertisement fragment is inserted between the two frames so as to achieve the goal of advertisement detection.
Description
Technical field
The invention provides a kind of method and related system that in a vision signal, detects advertising segment, refer in particular to a kind of method and related system that can carry out purposes of commercial detection with the high-effect relatively similarities and differences between the frame.
Background technology
The audio and video program service that is provided by wired or wireless broadcasting and TV medium is one of most important information source of advanced information society.Spectators can obtain useful news, knowledge, information or can relieve the audiovisual entertainment of body and mind from the audio and video program service.Yet under the consideration of commerce, the audio and video program regular meeting that the broadcasting and TV medium are provided has advertising segment to intert between normal program.Concerning spectators, these advertising segment regular meetings disturb the continuity of normal program, and spectators' time is also wasted in the puzzlement when causing spectators' rating normal program.When spectators will record the reference that be used as down in the future with normal program (maybe will record when playing after a while), these advertising segments more can expend the resource of user's recording video signal, and cause that the user can't retrieve easily and quickly, management, its vision signal of being recorded of access, and in routine techniques, existing technology lacks this vision signal advertising segment testing mechanism and system.
Summary of the invention
In general, because advertising segment is different with normal program, so before and after the insertion point of advertising segment, can cause the discontinuous of image frame, and existing broadcasting and TV medium can be after advertising segment finishes, it is a bit of that the preceding normal program of advertising segment is replayed, suitably to keep the continuity of normal program.Therefore, the invention provides a kind of method that can detect advertising segment in a vision signal, the difference of two consecutive frames in this vision signal relatively in regular turn is to find out the discontinuous part of image frame (having than the big-difference part between the picture of just adjacent two frames).Between discontinuous adjacent two frames of image frame, just might be the insertion point of advertising segment.After finding the discontinuous place of image frame,, just can search this given frame each frame (can be described as the forward direction search) before, to look for the frame whether identical with this frame (or similarity degree is quite big) is arranged at the discontinuous given frame of this picture.If at this given frame, can search the frame of identical (highly similar) really, just can assess out: this similar frame should be the advertising segment frame of normal program before, and this given frame should be exactly the frame of replaying behind this advertising segment.In other words, the frame of advertising segment should be just between the similar frame to it of this given frame.So, also just the place of advertising segment can have been detected.
As seen from the above description, the present invention only needs to carry out the search of forward direction at the frame of the discontinuous part of picture; If the picture of two consecutive frames is too big difference not, represent this two frame to belong to a series of gradual change frames in a series of dynamic images, can not the advertising segment insert division, just need not carry out the forward direction search yet.So, can significantly reduce the number of times that forward direction is searched.In addition, when a given frame being carried out the forward direction search and comparing, comprise and deposit this given frame frame before.Yet the present invention only need deposit the frame at the discontinuous place of picture, and only need will this given frame and the frame at the discontinuous place of picture relatively search mutually.Because if this given frame is the frame of replay after the advertising segment, the frame of identical with it (highly similar) must be before the advertising segment insert division, and that the advertising segment insert division just has picture is discontinuous.So the present invention only need deposit, searches, relatively have the discontinuous frame of picture to get final product before this given frame when a given frame being carried out the forward direction search.Simultaneously, the present invention need not deposit the pixel data of each pixel in each frame, can utilize an eigenvalue calculation rule of presetting that the feature extraction of one frame is the characteristic of correspondence Value Data; Carrying out forward direction at a given frame when searching, as long as deposit the characteristic of the discontinuous frame of picture, and characteristic that will this given frame and the characteristic comparison that each is deposited, whether similar frame is arranged with this given frame of relatively judging of characteristic.Above-mentioned all technical applications can both greatly be simplified the flow process that the present invention carries out purposes of commercial detection, reduce the system resource of use required for the present invention, make the present invention can be with the high-effect purposes of commercial detection of carrying out.
The present invention also provides a kind of signal processing system that can detect an advertising segment in a vision signal, comprise a difference comparison module, can be used in regular turn each frame of this vision signal being carried out a difference comparison step, and to set a relative different desired value, and if the difference degree of a present frame and a former frame during above a threshold value difference degree, with this present frame as a reference frame; And a search module, be used for searching a target frame that has this identical relative different desired value before this reference frame in the preset number frame, to determine the advertising segment position.In addition, this signal processing system still can comprise deposits module, is used to provide the required memory resource of this signal processing system, and the characteristic value of storing the former frame of the relative different mark value of each frame correspondence and each reference frame.This signal processing system still can comprise a similar comparison module, after searching target frame in search module, the characteristic value of the characteristic value of the former frame of this target frame and this reference frame relatively is with the similarity degree of the former frame content of judging this reference frame content and this target frame.On the other hand, this similar comparison module also can be used on this search module search when searching more than one target frame, be used for carrying out a similar comparison step, with the relatively characteristic value of these target frame and the similarity degree of this reference frame feature, the frame that will have maximum similarity degree then to the frame between the reference frame as advertising segment.
The present invention provide again a kind of can be in a vision signal recording-reproducing system of filtering one advertising segment, comprise a coding and decoding video unit, in order to this vision signal and the vision signal of filtering advertising segment of encoding of decoding: a video frame buffer unit, finished the vision signal of decoding in order to storage, and transmit this decoded video signal to one indicative control unit, be presented at a display unit when to be played; One purposes of commercial detection subsystem is in order to obtain and according to the characteristic value and the relative different mark value of this each frame of decoded video signal, to detect advertising segment; And an advertisement filtering subsystem, produce the vision signal of filtering advertising segment in order to testing result according to the purposes of commercial detection subsystem, deliver to this coding and decoding video unit then and encode, to treat subsequent access.
Utilize commercial detection method of the present invention and relevant signal processing system, the advertising segment that can will intert betwixt from a vision signal detects, make things convenient for that the user skips, the advertising segment in filtering or the clip video signal, allow the audio and video program service that the user can more effectively use, access broadcasting and TV medium provide.
Description of drawings
Fig. 1 is that advertising segment is inserted in schematic diagram between normal program in the exemplary video signal.
Fig. 2 shows an embodiment of the invention process algorithm.
Fig. 3 to Fig. 5 carries out the process schematic diagram of purposes of commercial detection in a vision signal for algorithm among Fig. 2.
Fig. 6 is the function block schematic diagram of signal processing system one embodiment of the present invention.
Fig. 7 is the function block schematic diagram of recording-reproducing system one embodiment of the present invention.
The reference numeral explanation
10, V: vision signal 12,14B, 16,18: program
14A, 14C: step 20: algorithm
30: signal processing system 32,461: difference comparison module
34,462: search module 36,463: deposit module
38,464: similar comparison module 40: recording-reproducing system
41: video input module 42: the coding and decoding video unit
44: video frame buffer unit 46: the purposes of commercial detection subsystem
48: advertisement filtering subsystem 50: indicative control unit
Sp1, Sp2: normal program Ss: paragraph fragment
Ad: advertising film F (a1)-F (k2), F (t1)-F (t6): frame
RPP (i) relative different mark
Embodiment
Please refer to Fig. 1.Fig. 1 is the schematic diagram of a vision signal 10.Vision signal 10 can provide different frames at different time in regular turn, as frame F (a1), F (a1+1) are provided in regular turn ... to F (a2), F (b1) to F (b2) up to frame F (k1) to F (k2) or the like, to present dynamic image.As discussed earlier, in the vision signal that the broadcasting and TV medium provide, meeting is interspersed advertising segment between normal program, and in the vision signal 10 of Fig. 1, promptly is interspersed with an advertising segment Ad between two fragment Sp1, Sp2 of normal program; Wherein, normal program fragment Sp1 have frame F (a1), F (a1+1) to F (a2), F (b1) to F (b2) ... up to F (c1) to F (c2) and F (d1) to F (d2); Have among the normal program fragment Sp2 frame F (h1) to F (h2), F (i1) to F (i2), F (j1) to F (j2) and F (k1) to F (k2) or the like, frame F (e1) then is used for presenting the dynamic image of advertising segment Ad to F (g2) up to frame F (g1) to F (e2), F (f1) to F (f2) or the like.
As is known to the person skilled in the art, can be with a series of dynamic image (as the dynamic image of the same camera lens of what is called) by presenting with a series of a plurality of frames with gradual change picture, no matter and be normal program or advertising segment, all be the set different series dynamic image to present its content.Picture is in the vision signal 10 of Fig. 1, and the frame F (a1) among the normal program fragment Sp1, F (a1+1) or the like are a series of frames with gradual change picture to frame F (a2), to present with a series of dynamic images.In like manner, frame F (b1) to F (b2), frame F (c1) to F (c2), frame F (d1) to F (d2), frame F (e1) to F (e2), frame F (f1) to F (f2), frame F (g1) to F (g2), frame F (h1) to F (h2), frame F (i1) to F (i2), frame F (j1) to F (j2) and even frame F (k1) to F (k2) or the like, all be respectively a series of frames, to present a series of dynamic image respectively with gradual change picture.
Between the gradual change frame with a series of dynamic images, the picture that two consecutive frames are presented also is gradual change, and it has the continuity on the picture image.As frame F (a1), F (a1+1) is that the picture that it presented also just has continuity with a series of gradual change frames; That is to say, frame F (a1) and F (a+1) though between variant, difference is little.With respect to similarity degree with a series of dynamic image gradual change interframe, will have between the frame of different series dynamic image big, than evident difference.For instance, when frame F (a1) is when being used for respectively presenting the dynamic image of two sections different series to F (a2) and frame F (b1) to F (b2), two consecutive frame F (a2), F (b1) just in time are located in the joining place of two different series images, and just have bigger difference between frame F (a2) and the F (b1), cause discontinuous on the picture.In like manner, if the frame F (e1) among the advertising segment Ad is when belonging to the dynamic image of different series to F (e2), frame F (f1) respectively to F (f2), consecutive frame F (e2) and F (f1) also have bigger difference, form discontinuous on the picture image.Certainly, advertising segment and normal program all are made up of the dynamic image of different series respectively, so locating of being connected with normal program of advertising segment also must have the discontinuous of picture.Picture forms discontinuously in Fig. 1 with regard to having bigger difference between consecutive frame F (d2), the F (e1), also have discontinuous on the image frame between frame F (g2), the F (h1).In other words, the insertion point of advertising segment Ad must present the discontinuous of image frame.
In addition, present broadcasting and TV medium also can be after advertising segment finishes, and it is a bit of the normal program before the advertising segment to be replayed again, and makes things convenient for spectators to be connected the programme content of advertising segment front and back.In Fig. 1, just represent this paragraph fragment to the paragraph fragment Ss that frame F (d2) is concatenated into by frame F (c1).After advertising segment Ad/ finishes, will repeat to broadcast the paragraph fragment Ss that originally in normal program Sp1, had occurred among the normal program fragment Sp2.In other words, frame F (h1) is to the image content of F (i2), and is just identical to F (d2) with frame F (c1) respectively.
By above-mentioned discussion as can be known, the interframe before and after the advertising segment has two kinds of features; At first, the beginning and end of advertising segment has the discontinuous of picture.Secondly, can the frame that repeat be arranged because of the replay of paragraph fragment in the normal program before and after the advertising segment.The present invention utilizes these two kinds of features to carry out purposes of commercial detection.Please refer to Fig. 2.Fig. 2 promptly presents the embodiment that the technology of the present invention is implemented with the form of algorithm.
Shown in the algorithm among Fig. 2 20, when the present invention will carry out purposes of commercial detection at a vision signal V, can carry out program 12 to each frame among this vision signal V in regular turn.When i the frame F (i) among the vision signal V carried out program 12, can carry out step 14A earlier, with the difference degree of comparison frame F (i) with its former frame F (i-1), and with the relative different mark RPP (i) of diversity ratio outcome record in a correspondence.If the difference degree between these two consecutive frame F (i) and the F (i-1) is quite big, discontinuous on the picture arranged between representative frame F (i-1) and the F (i), just the value of relative different mark RPP (i) can be made as " 1 " at this moment.Otherwise,, just the value of relative different mark RPP (i) can be made as " 0 " if the difference between frame F (i-1) and the F (i) is little.When this step of specific implementation, can calculate a corresponding characteristic value data E (i) and E (i-1) respectively according to frame F (i) and picture frame F (i-1), if the difference between characteristic value data E (i) and the E (i-1) has surpassed a threshold value difference degree, relative different mark RPP (i) can be made as " 1 ", otherwise then be " 0 ".Wherein, characteristic value data E (i) can be all pixel datas among the frame F (i) (as the brightness or the color data of all pixels) and, or the distribution scenario (histogram) of brightness or color among the frame F (i), or even be used as characteristic value data E (i) through certain frequency domain conversion (as two-dimentional cosine conversion) resulting result in back according to frame F (i).Basically, a frame characteristic of correspondence Value Data, the size of its data volume can be less than the size of all phase prime number certificates of this frame itself, to accelerate the carrying out of step 14A.
The diversity ratio of finishing step 14A after, algorithm 20 will continue to program 14B.In program 14B,, then picture frame F (i) is considered as one with reference to picture frame, and in step 14C, its characteristic of correspondence Value Data (as aforementioned characteristic value data E (i-1)) is noted if the value of relative different mark RPP (i) is " 1 "; Simultaneously, also can begin to carry out program 16, to carry out the forward direction search.When the value of relative different mark RPP (i) is " 1 ", between representative frame F (i-1) and the F (i) the discontinuous of picture arranged, might be the insertion point of advertisement, but also might only be the place that the different series dynamic image is connected.For further affirmation, just can carry out the forward direction search of program 16, to search frame F (i) each frame before, look at whether to have frame similar to frame F (i).If a certain frame F (j) identical with frame F (i) (or very similar) arranged, belong to a series of dynamic images with regard to representative frame F (i) and frame F (j), and order after frame F (i) just belong to the replay fragment of this dynamic image; So, just can infer: advertising segment should be just between frame F (j) and frame F (i).
As shown in Figure 2, when the forward direction that carries out program 16 is searched, the scope that forward direction is searched be can configure earlier, two integer M, N just set.Then, just frame F (i-N) can be considered as carrying out the frame that forward direction is searched to F (i-M-N), with to frame F (i-N), F (i-N-1) until j the frame F (j) between frame F (i-M-N) carries out program 18.In program 18, can mark whether to be " 1 " by the corresponding relative different of first judgment frame F (j); If " 1 " carries out a similar comparison step again to compare the similarity degree between frame F (j-1) and the F (i); If both similarity degrees are high, representative frame F (i) is the replay of frame F (j-1), and advertising segment should be located between frame F (j-1) and the frame F (i).Relative, if all can not find the similar frame with frame F (i) in the frame of proper range, originally the picture between frame F (i), F (i-1) was discontinuous may be that the different series dynamic image causes in representative, was not the insertion point of advertising segment.
Carry out similarity degree in the program 18 relatively the time actual, equally also can according to frame F (j-1) and frame F (i) institute respectively the characteristic of correspondence Value Data come comparison; If the difference of two frame character pair Value Datas is less than a threshold trait difference, in the equivalence with regard to the similarity degree between representative frame F (j-1) and the frame F (i) greater than a threshold value similarity degree, frame F (j-1) and F (i) can be considered as similar, belong to frame with a series of dynamic images.
In addition, also can find out in the program 18 by algorithm 20, the present invention do not need to each frame in the forward direction search area (just frame F (i-M) is to F (i-M-N)) carry out in the program 18 similarity degree relatively, as long as the frame F (j-1) of correspondence is carried out the comparison of similarity degree with frame F (i) during for " 1 " at relative different mark RPP (j).Because the purpose of forward direction search is to check whether frame F (i) is the frame of replaying behind advertising segment, basically as long as frame F (i) is compared with advertising segment frame before; Though not knowing which frame when carrying out program 18 is the frame before advertising segment, the frame before the frame before the advertising segment must also be the discontinuous place of picture.So, carrying out forward direction when searching, only the previous frame at frame F (i) and the discontinuous place of picture need be compared and get final product.Because relative different mark RPP (j) is used for representing whether have picture discontinuous between frame F (j-1) and the frame F (j), so as long as step 18 is just carried out the comparison of similarity degree to frame F (i) and frame F (j-1) (the previous frame at the discontinuous place of picture just) just when relative different mark RPP (j) is " 1 ".
In addition, when carrying out the forward direction search at frame F (i), need record the correlated characteristic Value Data of frame F (j), itself and frame F (i) could be compared.But by principle discussed above as can be known, in preferred embodiment of the present invention, also as long as the characteristic value data of the previous frame at the discontinuous place of record picture just can successfully carry out the forward direction search.And this purpose of step 14C just, it is after confirming that the corresponding relative different mark RPP (i) of a frame F (i) be " 1 ", ability is noted the characteristic value data of frame F (i-1).In addition, the size of data of a frame characteristic value data is original just less than the size of data of all pixel datas of this frame itself; More than these factors all make the present invention to realize in the mode of quick, high-effect, low system requirements.
For further specifying the situation that algorithm 20 of the present invention is implemented, please continue with reference to figure 3 to Fig. 5 (and simultaneously with reference to figure 2).Fig. 3 to Fig. 5 signal be the situation of the present invention when in a vision signal V, carrying out purposes of commercial detection.At first, as shown in Figure 3, suppose that algorithm 20 proceeds to t1 frame among the vision signal V in regular turn, just make the i=t1 among Fig. 2, and begin to carry out program 12.Therefore, algorithm 20 begins the difference between comparison frame F (t1) and the F (t1-1) in step 14A, to calculate corresponding relative different mark RPP (t1).Just as discussed earlier, step 14A can according to frame F (t1-1) and frame F (t1) distinguish characteristic of correspondence Value Data E (t1-1) and E (t1) carries out the comparison of difference.In Fig. 3, suppose that frame F (t1) and F (t1-1) belong to two gradual change frames with a series of dynamic images, so difference between the two is little, and make corresponding relative different mark RPP (t1) be " 0 ".Because relative different mark RPP (t1) is " 0 ", the neither need of all programs behind the step 14A carry out, and also needn't continue recording feature Value Data E (t1-1), and the memory resource that it can be taken discharges.Next, algorithm 20 just can continue to a time frame F (t1+1) (just making i=t1+1).
The example of continuity Fig. 3 in Fig. 4, supposes that algorithm 20 has proceeded to frame F (t2), just makes the i=t2 among Fig. 2, and begins to carry out program 12.Herein, suppose that frame F (t2) and frame F (t2-1) belong to the dynamic image of different series, so being arranged, suitable difference forms the discontinuous of picture between the two, make that also corresponding relative different mark RPP (t2) is " 1 ", and algorithm 20 also can continue to program 14B after step 14A.In step 14C, it is a reference frame that frame F (t2) can be taken as, and characteristic value data E (t2-1) can go on record, and begins to carry out the forward direction search in the program 16.And in program 16, can begin to check the relative different mark RPP of each frame correspondence to F (t2-N) to frame F (t2-M-N).In the example of Fig. 4, suppose that the relative different mark of these frame correspondences is all " 0 ", just relative different mark RPP (t2-M-N) is all " 0 " to RPP (t2-N), and then algorithm 20 also compares with regard to the similarity degree that need not carry out in the program 18.After the program 12 that finishes frame F (t2) is carried out when algorithm, can remain with frame F (t2-1) characteristic of correspondence Value Data E (t2-1).Next, algorithm 20 will continue to a time frame F (t2+1).
The example of continuity Fig. 3, Fig. 4 in Fig. 5, supposes that algorithm 20 has proceeded to frame F (t6), just makes the i=t6 among Fig. 2 and begins to carry out program 12.In the example of Fig. 5, suppose that frame F (t6-1) has suitable difference with F (t6) and makes corresponding relative different mark RPP (t6) be " 1 ", and algorithm 20 also just begins to proceed to program 14B.When the forward direction in carrying out program 16 is searched, just can check respectively that to frame F (t6-N) relative different of each frame correspondence marks whether to be " 0 " to frame F (t6-M-N).In the example of Fig. 5, suppose in the scope of target frame have frame F (t3), F (t4) and F (t5) pairing relative different mark RPP (t3), RPP (t4) and RPP (t5) to be " 1 ".So algorithm 20 will be in program 18, the similarity degree that j=t3, t4 and t5 are carried out in the program 18 compares, to compare the difference between frame F (t3-1), F (t4-1) and F (t5-1) and the frame F (t6) respectively.Please note, because relative different mark RPP (t3), RPP (t4) and RPP (t5) are all " 1 ", so frame F (t3-1), F (t4-1) and F (t5-1) distinguish characteristic of correspondence Value Data E (t3-1), E (t4-1) and E (t5-1) goes on record all in advance, and when carrying out program 18, just can utilize frame F (t6) characteristic of correspondence Value Data E (t6) to come to compare with characteristic value data E (t3-1), E (t4-1) and E (t5-1) respectively, to compare the similarity degree of frame F (t6) and these interframe.Suppose frame F (t6) and frame F (t4-1) similarity degree maximum, just can infer that advertising segment is to be inserted between frame F (t4-1) and the F (t6), reaches the purpose of purposes of commercial detection of the present invention.
Please refer to Fig. 6 (and in the lump with reference to figure 2).Algorithm 20 among Fig. 2 can realize with the signal processing system among Fig. 6 30.Signal processing system 30 comprises a difference comparison module 32, a search module 34.Difference comparison module 32 is used for to perform calculations step 14A (Fig. 2) in the method 20 of vision signal V; In the time will carrying out the forward direction search, search module 34 can be carried out program 16.Signal processing system 30 still can comprise similar comparison module 38, is used for receiving the output of depositing module 36, with the similarity degree comparison of executive program 18.Deposit module 36 required memory resource when being used for supporting above-mentioned each module work, as the relative different desired value of depositing each frame, but and performing step 14C, in the characteristic value that writes down associated frame whenever necessary.Signal processing system 30 can be built and be placed an Equipment for recording and playing back images, as the video recorder of video tape, CD, hard disk formula, so that carry out purposes of commercial detection in the vision signal V that it is noted.In addition, each module in the signal processing system 30 can realize with firmware or hardware respectively.For instance, the function of difference comparison module 34, search module 36, similar comparison module 38 can and realize by a processor different firmware code of execution, can be realized by memory (as random access memory) and deposit module 36.In addition, signal processing system 30 also can be built and be placed a multimedia computer.Present multimedia computer can come stored video signal with hard disk or CD, and signal processing system 30 promptly can be used to detect the advertising segment in these vision signals.For instance, can be by obtaining vision signal V in hard disk or the CD, the function of difference comparison module 32, search module 34 and similar comparison module 38 can be realized by the central processing unit (or the treatment circuit on the display card) of the computer mode with the software program for execution code.
Carry out forward direction when searching in algorithm 20 of the present invention (Fig. 2), the scope (setting of M, N just) of waiting to search frame can be looked actual conditions and be set.For instance, if the advertising segment that the broadcasting and TV medium insert can not be shorter than 30 seconds, and per second is corresponding to 60 frames, and the N among Fig. 2 just can be set at 1800 (promptly 30 multiply by 60).In addition, the advertising segment that the broadcasting and TV medium insert is the upper limit of the free length of meeting (for example the advertising segment of a joint was no longer than 5 minutes) usually, and the M of forward direction search just can set in view of the above.According to similar principle, also can determine to deposit in the signal processing system 30 the required memory resource size of module 36.Certainly, the M that forward direction is searched in the algorithm 20 also can be parameter (as with the different parameters that change of i) with N.
Fig. 7 is the function side of the recording-reproducing system one embodiment of the present invention schematic diagram of determining.Wherein, recording-reproducing system 40 can be a magnetic tape type videocorder, optical-disk type videocorder, hard disk formula videocorder, or utilizes hard disk or CD to store a multimedia computer of this vision signal.Recording-reproducing system 40 comprises a coding and decoding video unit 42, in order to decoded video signal and coding one vision signal of filtering advertising segment; One video frame buffer unit 44 in order to storing a decoded vision signal, and then is sent to an indicative control unit 50; One purposes of commercial detection subsystem 46 is in order to obtain and according to the characteristic value and the relative different mark value of this each frame of decoded video signal, to detect advertising segment; And an advertisement filtering subsystem 48, in order to according to the testing result of purposes of commercial detection subsystem 46 producing the vision signal of filtering advertising segment, and deliver to coding and decoding video unit 42 and encode.This recording-reproducing system 40 still can comprise a video input module 41, is used for receiving video signals, decodes for coding and decoding video unit 42, and wherein this vision signal can provide frame to present dynamic image according to a preset order.In addition, purposes of commercial detection subsystem 46 can comprise a difference comparison module 461, in regular turn each frame of decoded video signal being carried out a difference comparison step, and to set a relative different desired value and a reference frame; And a search module 462, in order to search the target frame that has identical relative different desired value before the reference frame in the preset number frame, to determine advertising segment.Purposes of commercial detection subsystem 46 still can comprise deposits module 463, in order to the characteristic value of the former frame of the relative different mark value of storing each frame correspondence and each reference frame; And a similar comparison module 464, to carry out similarity degree relatively.One carry at this, and each module in the purposes of commercial detection subsystem 46 can realize with firmware or hardware respectively.
In routine techniques,, cause the user can't use the vision signal that it provided more efficiently because can't effectively detect the advertising segment that interts in the vision signal.In comparison, the feature that the present invention can utilize advertising segment to insert front and back is carried out purposes of commercial detection, and can realize with calculating, the memory resource of simplifying, the user skips to assist, the advertising segment in the filtering vision signal, makes things convenient for the service of user's access more efficiently, arrangement, utilization vision signal.
The above only is preferred embodiment of the present invention, and all equivalences of carrying out according to claim of the present invention change and revise, and all should belong to covering scope of the present invention.
Claims (9)
1. method that can detect an advertising segment in a vision signal comprises:
In regular turn each frame of this vision signal is carried out a difference comparison step, setting a relative different desired value of each frame, and surpass this present frame of a threshold value difference degree as a reference frame with the difference degree of a present frame and a former frame; And
Search a target frame that has the relative different desired value identical before this reference frame in the preset number frame, to determine this advertising segment position with this reference frame.
2. the method for claim 1 also comprises:
After searching this target frame, the characteristic value of the characteristic value of the former frame of this target frame and this reference frame relatively is with the similarity degree of the former frame content of judging this reference frame content and this target frame.
3. the method for claim 1 also comprises:
Search more than one this target frame; And
Carry out a similar comparison step, with the relatively characteristic value of these a plurality of these target frame and the similarity degree of this reference frame feature.
4. method as claimed in claim 3, if similarity degree is greater than a threshold value similarity degree, select one to have peaked this target frame of similarity degree, and the frame of judging this advertising segment is arranged in this and have the former frame of peaked this target frame of similarity degree to this reference frame.
5. the signal processing system that can detect an advertising segment in a vision signal can be built and be placed an Equipment for recording and playing back images or a multimedia computer, comprising:
One difference comparison module, in order in regular turn each frame of this vision signal is carried out a difference comparison step, setting a relative different desired value of each frame, and surpass this present frame of a threshold value difference degree as a reference frame with the difference degree of a present frame and a former frame; And
One search module is in order to search a target frame that has the relative different desired value identical with this reference frame before this reference frame in the preset number frame, to determine this advertising segment position.
6. signal processing system as claimed in claim 5 also comprises:
One similar comparison module, after searching this target frame in this search module, the characteristic value of the characteristic value of the former frame of this target frame and this reference frame relatively is with the similarity degree of the former frame content of judging this reference frame content and this target frame.
7. signal processing system as claimed in claim 5 also comprises:
One similar comparison module in order to search more than one this target frame in this search module search after, carries out a similar comparison step with the relatively characteristic value of these a plurality of these target frame and the similarity degree of this reference frame feature.
8. signal processing system as claimed in claim 5 comprises:
One deposits module, in order to providing this signal processing system required memory resource, and stores the relative different desired value of each frame correspondence and the characteristic value of the former frame of this reference frame respectively.
One kind can be in a vision signal recording-reproducing system of filtering one advertising segment, comprising:
One coding and decoding video unit is in order to this vision signal and the coding one this vision signal of this advertising segment of filtering of decoding;
One video frame buffer unit in order to storing decoded this vision signal, and transmits this decoded video signal to one indicative control unit;
One purposes of commercial detection subsystem, week is to obtain and according to the characteristic value and the relative different desired value of this each frame of decoded video signal, to detect this advertising segment; And
One advertisement filtering subsystem, in order to according to the testing result of this purposes of commercial detection subsystem producing this this vision signal of this advertising segment of filtering, and deliver to this coding and decoding video unit and encode;
Wherein this purposes of commercial detection subsystem comprises:
One difference comparison module, in order in regular turn each frame of this decoded video signal is carried out a difference comparison step, setting a relative different desired value of each frame, and surpass this present frame of a threshold value difference degree as a reference frame with the difference degree of a present frame and a former frame; And
One search module is in order to search a target frame that has the relative different desired value identical with this reference frame before this reference frame in the preset number frame, to determine this advertising segment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100558090A CN1327701C (en) | 2004-08-04 | 2004-08-04 | Method and relative system for high efficiency ad detection in video signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100558090A CN1327701C (en) | 2004-08-04 | 2004-08-04 | Method and relative system for high efficiency ad detection in video signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1589002A CN1589002A (en) | 2005-03-02 |
| CN1327701C true CN1327701C (en) | 2007-07-18 |
Family
ID=34603203
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004100558090A Active CN1327701C (en) | 2004-08-04 | 2004-08-04 | Method and relative system for high efficiency ad detection in video signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1327701C (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100580693C (en) * | 2008-01-30 | 2010-01-13 | 中国科学院计算技术研究所 | Advertisement detecting and recognizing method and system |
| CN101576955B (en) * | 2009-06-22 | 2011-10-05 | 中国科学院计算技术研究所 | Method and system for detecting advertisement in audio/video |
| CN105103566B (en) * | 2013-03-15 | 2019-05-21 | 构造数据有限责任公司 | Video clip is for identification so as to the system and method that show contextual content |
| CN105763884A (en) * | 2014-12-18 | 2016-07-13 | 广州市动景计算机科技有限公司 | Video processing method, device and apparatus |
| CN105791887A (en) * | 2014-12-23 | 2016-07-20 | 腾讯科技（北京）有限公司 | Information processing method, client and server |
| CN105337793B (en) * | 2015-09-24 | 2018-08-07 | 网宿科技股份有限公司 | A kind of network line broadcast monitoring method, system and a kind of server |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1130004A (en) * | 1993-08-31 | 1996-08-28 | 联合企业股份有限公司和无线电产业保护两合公司 | Process and device for detecting undesirable video scenes |
| WO2001035409A2 (en) * | 1999-11-10 | 2001-05-17 | Thomson Licensing S.A. | Commercial skip and chapter delineation feature on recordable media |
| CN1543096A (en) * | 2003-04-30 | 2004-11-03 | ���µ�����ҵ��ʽ���� | Equipment and method for automatic detection of television advertisement |
-
2004
- 2004-08-04 CN CNB2004100558090A patent/CN1327701C/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1130004A (en) * | 1993-08-31 | 1996-08-28 | 联合企业股份有限公司和无线电产业保护两合公司 | Process and device for detecting undesirable video scenes |
| WO2001035409A2 (en) * | 1999-11-10 | 2001-05-17 | Thomson Licensing S.A. | Commercial skip and chapter delineation feature on recordable media |
| CN1543096A (en) * | 2003-04-30 | 2004-11-03 | ���µ�����ҵ��ʽ���� | Equipment and method for automatic detection of television advertisement |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1589002A (en) | 2005-03-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11070871B2 (en) | Methods and apparatus to detect commercial advertisements associated with media presentations | |
| Chen et al. | Video sequence matching based on temporal ordinal measurement | |
| JP5687772B2 (en) | Dynamic content insertion using content signatures | |
| US8671109B2 (en) | Content-based video copy detection | |
| US8978060B2 (en) | Systems, methods, and media for presenting advertisements | |
| US9002175B1 (en) | Automated video trailer creation | |
| US20080270373A1 (en) | Method and Apparatus for Content Item Signature Matching | |
| US10841637B2 (en) | Time-adapted content delivery system and method | |
| TWI242376B (en) | Method and related system for detecting advertising by integrating results based on different detecting rules | |
| TWI245568B (en) | Method and related system for high efficient advertising detection | |
| US20090320060A1 (en) | Advertisement signature tracking | |
| US20090320063A1 (en) | Local advertisement insertion detection | |
| US11689753B2 (en) | Use of steganographically-encoded data as basis to disambiguate fingerprint-based channel-multi-match | |
| JP2007524321A (en) | Video trailer | |
| KR102390142B1 (en) | Synchronizing ancillary data to content including audio | |
| CN104683852A (en) | Method and equipment for processing broadcasting information | |
| CN1327701C (en) | Method and relative system for high efficiency ad detection in video signal | |
| CN100551014C (en) | The method of contents processing apparatus, contents processing | |
| KR20200138710A (en) | Coordinates as auxiliary data | |
| Yeh et al. | TV commercial detection in news program videos | |
| US20120010736A1 (en) | Spotting multimedia | |
| US20220076708A1 (en) | Video editing | |
| US10951935B2 (en) | Media environment driven content distribution platform | |
| US20130151544A1 (en) | Information processing apparatus, information processing method, and progam | |
| CN1320822C (en) | Method and relative system for cross detecting ad fragment using different detection principle |
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 |