WO2009000204A1

WO2009000204A1 - A method and a system of adding advertisement information into a media stream

Info

Publication number: WO2009000204A1
Application number: PCT/CN2008/071426
Authority: WO
Inventors: Zhiping Meng
Original assignee: Zhiping Meng
Priority date: 2007-06-25
Filing date: 2008-06-24
Publication date: 2008-12-31
Also published as: CN101072340B; CN101072340A

Abstract

A method and a system of adding advertisement information into a media stream are proposed. Usually, there is relativity between the advertisement information and the content or transmission objects or transmission time and place of the media stream, so it is more easily for clients to receive the content of advertisement or auxiliary information. The invention needs to process media stream files based on content, and add more advertisement information, then transmit the processed media files to clients through the network. When adding the advertisement information, firstly present the current advertisement searching information and form advertisement category storage information, and then extract the video, audio and text of the media stream to recognize. If it is found that the advertisement content is matching with the recognition result, add the advertisement information into the media stream to form an advertisement description file which is finally transmitted to clients together with the corresponding media stream. Analyzing the advertisement description file, the clients retrieve the advertisement information from the advertisement storage server and play it.

Description

Method and system for adding advertising information in streaming media

Embodiments of the present invention relate to the field of communications, and in particular, to a method for adding advertisement information to a streaming media, the advertisement information including an advertisement, introduction information about video and audio content, and the like. Background technique

When advertised, it's best to target the ad to the appropriate prospects rather than selectively advertising to the general public. If it is known that a certain user is watching a specific television content, it can be assumed that the user is interested in the part of the audiovisual content, and it can be inferred that the user is likely to be interested in the related content of the part of the audiovisual content. Under this hypothetical relationship, an advertisement related to the user pushing and playing the content of the audiovisual program is undoubtedly the form of advertisement that the user is most likely to receive.

Although this concept is reflected in existing analog and digital TVs, there is a lack of automatic and quick means to accomplish this complex task. With the widespread use of the Internet, advertisers have used the Internet to promote their products and services as they currently promote their products and services on television. On a particular Web site or video site, although there are many ads that target the site, most of the headlines on the site are untargeted, so they are inefficient and often less effective. Advertising on video sites has nothing to do with video content, and video ads are usually inserted in one end.

U.S. Patent No. 5,937,392 issued to Alberts entitled "Title Advertising Display System and Advertising Frequency Control Method" attempts to specifically target users seeking certain types of information to alleviate untargeted advertising problems. For example, when a user searches for a "computer" on a yellow page or on a search engine, the user can receive targeted advertisements from a computer manufacturer, as well as advertisements that can be targeted to users in a particular geographic area, and the like.

U.S. Patent No. 5,806,061 issued to Merriman et al., entitled "Method of Sending, Aligning, and Measuring Advertisements Through the Internet", discloses a A device in which individual users perform data tracking to locate advertisements.

U.S. Patent No. 5,974,451, to Simmons, entitled "Method for Assigning a Billing System to an External Computer Accessing a Wide Area Computer Network", advertises via the Internet based on a particular request by a user for information.

U.S. Patent No. 59,334,811 to Angles et al., entitled "System and Method for Sending Customized Advertisements in an Interactive Communication Network", delivers customized advertisements to users over the Internet based on the user's spending power or characteristics.

Finally, the Chinese patent No. CN00817788 titled "Selective Media Stream Advertising Technology", a real-time advertisement adding system based on video and audio content, generates an advertisement page by adding recognition to voice, video and text, and adds Forwarded to the user in the data stream. This advertising process is a process of discovery and release. It is a real-time advertising system. The matching advertising information is directly released to users without being stored and processed. This requires the advertising system to have strong information recognition and processing capabilities, and requires a strong database and algorithm support. When the advertising resources continue to expand, the system will be difficult to support a large amount of computing. If the advertisement is not delivered in time, the value of the advertisement associated with the content is lost. Moreover, the advertisement information of the patent is sent to the user side by being loaded into the media stream, which often requires the support of the media transmission protocol, and does not require the separation of content and services. Moreover, as a function of image recognition, since the recognition of video images is still in the research stage, no product can guarantee the recognition function of any object, so the patent is difficult to apply in reality. And the patent only considers the broadcast media as a means of playing media, without considering the situation of on-demand. Summary of the invention

The embodiment of the invention provides a method and a system for adding advertisement information in a streaming media, so as to solve the problem that the advertisement cannot be designed according to the relevance of the potential content and the advertisement required by the user in the prior art.

The purpose of the present invention is to load advertisement content or advertisement content links onto existing media files, which are usually stored in a streaming media server or a proprietary storage device, such as video and audio files or Flash. Can be streamed document. When adding an advertisement, it is considered to associate the advertisement with the content of the existing media file, so the existing media file needs to be associated, for example, the user can launch the advertisement of the jersey and the sneaker when watching the football match. The steps to be processed are as follows:

First, sorting out all the advertising information, abstracting each advertising information into key advertising keywords, advertising voice recordings, image and trademarks of the products in the advertisements, and so on. For example, Coca-Cola's advertisement corresponds to the catalogue of food and beverage soft drinks. The advertising keyword is Coca-Cola (or: Coca-Cola), and the pronunciation of Coca-Cola (such as Chinese, English or other regional language pronunciations) is recorded. Image of Coca-Cola's trademark and its typical package shape. Similar items such as Pepsi, Coca-Cola, etc., which are similar to Coca-Cola, are stored in the same soft drink catalog. Such ad items can also be expanded, for example, the special ad word "I can" or any information related to the ad item can be added to the ad item according to actual needs, which can improve the search for existing streaming media file content to seek a match. , to expand the hit rate of matching, thereby adding more advertising information, but also increasing the advertising value of the streaming media itself. The above operations are usually done by means of a database system. Finally, an advertisement item usually contains one or more of an image part, a voice part, and a text part. Most of the information in the advertisement item is for searching and/or adding a media advertisement. When the information is similarly determined, the link information of the advertisement (link to the real advertisement position) or the real advertisement content for the advertisement product may be retained in the advertisement information, including: video and audio, animation, image, text, and the like.

Second, after having such advertising classification information, it is necessary to process the existing streaming media files. When processing streaming media files, using the content analysis module, the module will separate three streams from the streaming media file, namely: video stream, audio stream and character stream, and then analyze the three streams, using manual identification. The image automatic recognition and the like are compared with the image information in the search information column in the advertisement information, and the advertisement information is added in the appropriate position of the video in the streaming media file, and the joining method is to generate the advertisement description file (profile) of the streaming media. „ 类似 In a similar way, the content analysis module is used to analyze the audio stream and the character stream, and the advertisement information is added. The information of the advertisement mainly refers to the content position of the advertisement, which may be the physical location of the specific content in the storage medium. It can also be an abstract location in a network. Advertising information can be any form of media content. Finally, the streaming media file is sent to the user along with the generated advertisement description file. The user plays the streaming media file through the playing device, and reads the advertisement description file of the played streaming media through software, and plays the advertisement according to the description of the advertisement description file.

Compared with the prior art, the embodiment of the invention has the following advantages:

The form of advertising is more user-friendly, and the advertisement may be equal to what the user needs for the user. Moreover, the effect of the advertisement is greatly improved, and the advertisement can be more accurately placed on the user according to the behavior of the user to watch the advertisement;

In the embodiment of the present invention, an advertisement is designed according to the relevance of the potential content and the advertisement required by the user, and the advertisement form related to the user's push and the content of the video and audio program being played is used, and the user can be accurately determined according to the past behavior of the user. Ad serving. DRAWINGS

The above description of the invention will be better understood from the following detailed description of the embodiments of the invention. The above and the following text or illustrations are intended to be illustrative of the exemplary embodiments of the present invention.

1 is a schematic structural view of a system of the present invention;

2 is a schematic diagram of generating and using an advertisement description file according to the present invention;

3 is a flow chart of the present invention for adding an advertisement to a streaming media file and transmitting it to a client; FIG. 4 is a schematic structural diagram of an advertisement classification storage information according to the present invention;

Figure 5 is a schematic diagram of the principle of the content analysis module of the present invention;

6 is a schematic flowchart of the internal operation of the user playing device according to the present invention;

7 is a schematic diagram of an internal process of an advertisement storage server according to the present invention;

FIG. 8 is a schematic diagram of a flow of expanding advertisement search information in an advertisement storage server according to the present invention; FIG.

9 is a schematic diagram of a play window of a user playback device according to the present invention;

FIG. 10 is a schematic diagram of a user request index according to the present invention. detailed description

The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

The basic idea of the present invention is: When an advertisement is added to a media file such as video or audio, an advertisement classification storage information is created by a method of classifying all advertisements and extracting information. Then use manual recognition or automatic recognition or both methods to identify the image, sound, and text parts of the media file, and match the identified result with the existing advertisement classification information to generate the successfully matched part. An advertisement description file that records information such as where the advertisement was taken, when it was played, or during certain frames, and the area of the advertisement. Finally, the advertisement description file and the streaming media file are transmitted to the user playing device, and when the streaming media file is played, the user extracts the advertisement description information through the software, and plays the advertisement related to the content at the set position and time.

As shown in Fig. 1, a schematic structural diagram of the system of the present invention, this embodiment describes a possible system structure of the present invention. The function of the streaming media storage server 110 is to store a streaming media file, which is a file structure that can be transmitted by means of a stream, but is not limited to the streaming mode, because the system is not a real-time system, that is, it must be an advertisement description file. Then, the user can be activated to join the advertisement at the time of playing, so the media file stored in the non-streaming manner is also applicable to the present invention. These common streaming media files include: WMV (Windows Media Video, Windows Media Video) / Mpeg (Moving Picture Expert Group) 2/Mpeg4/H.264 or more advanced encoding formats. The content analysis module or server 120 may be a separate device or a functional module that is placed together with the streaming media file. Except for Figure 1, the content analysis module is used to express the information, and the main function is to store the information 210 and the streaming media file in combination with the advertisement classification. The content itself produces an ad description file 230 (210 and 230 are not shown in Figure 1, which should be in Figure 2). The advertisement classification storage information is generated by the advertisement storage server 140, and the advertisement storage server 140 is configured to store all the advertisement information. The push server 130 pushes the streaming media file and the advertisement description file corresponding to the streaming media file to the user playback device via the network transmission 150. The push server 130 can push the streaming media file and the advertisement description file 230 of the streaming media file 220, respectively, or push them at the same time. At the same time, push Yueliang Server 130 is an optional device that may be replaced by a streaming server such as a streaming storage server. It should be noted that the streaming media file and the corresponding advertising description file corresponding to the streaming media file are finally pushed to the user playing device. The network transmission 150 can be any network that transmits media files, such as a network structure based on IP or ATM (Asynchronous Transfer Mode). The user playback device 190 includes a PC 160, a mobile terminal 170, a fixed terminal 180, and the like. The user media playback device is constantly changing and merging, and thus the user playback device is not limited thereto, as long as the connection can be performed through a network (including wired or wireless). The devices of the media files are all in the playback device. The fixed terminal 180 refers to a mode in which a streaming media file is played through a set top box plus a television. When playing the streaming media file, the user playback device 190 needs to parse the advertisement description file 230 at the same time, and then retrieve the advertisement information stored in the advertisement storage server 140 by the content of the advertisement description file. The advertisement storage server is not limited to a specific server, but any service device that can provide advertisements to users. Of course, in this case, the advertisement classification storage information 210 must reflect the advertisement information of all advertisement storage servers.

In FIG. 2, the streaming media file 220 is first stored in the streaming media storage server 110, the advertisement classification storage information 210 is generated by the advertisement storage server 140, and the advertisement classification storage information 210 is stored in the content analysis module 120 through the streaming media. The storage server and the content analysis module cooperate to generate an advertisement description file 230, and the advertisement description file and the streaming media file that generates the advertisement description file are correspondingly sent to the user playback device 190. After processing by the content analysis module, each streaming media file corresponds to an advertisement description file, or a streaming media file may correspond to multiple advertisement description files, because different users may have different advertisement requirements, Configuring different ad description files for different users will improve the user's viewing of the ads. However, in the embodiment of the present invention, a method for generating an advertisement description file by content correlation is mainly discussed. Therefore, a method in which one media file corresponds to a plurality of advertisement description files is not within the scope of the present invention. In some special cases, the advertisement storage server 140 that the user retrieves the advertisement information 240 and the advertisement storage server that generates the advertisement classification storage information are not the same server, especially when the advertisement is stored using distributed storage or mirroring technology. In practice, the content analysis module may appear as a form of content analysis server, or as The content analysis module appears in a server that processes the contents of the streaming media file, such as a streaming media storage server or a push server. The advertisement description file generated by the content analysis module can be stored on the same server as the streaming media file or stored by a dedicated advertisement description server.

As shown in FIG. 3, this figure is a flow chart of the present invention for adding an advertisement to a streaming media file and sending it to the client. Includes the following steps:

310. When processing the streaming media file, firstly, the streaming media file needs to be opened, and the opening may be opened by directly opening the streaming media storage server 110 or opening the streaming media file through the client/server.

320, extracting video/audio/character files. Among them, there are two methods for extracting video, audio, and text. The first is to extract the video, audio, and text information in the media file to form three separate files, including video files, audio files, character files, and then The three files are processed separately; the second is to extract the video stream, the audio stream and the character stream from the streaming media file, and then process the three streams separately. The three files or three streams extracted from the streaming media file are processed by the content analysis module 120.

330. Determine whether the video, audio, and characters extracted in the streaming media file match the advertisement information. If yes, go to step 340; if otherwise, continue to analyze whether there are videos, audios, and characters that match the advertisement information.

340. Add an advertisement information to the advertisement description file.

350. Finally, the advertisement description file is sent to the client for parsing, and this process usually occurs along with the behavior of the streaming media file sent to the client. After obtaining the advertisement description file, the client parses the advertisement description file through the advertisement description parser, and retrieves the advertisement information of the advertisement storage server from the content of the advertisement description file, and then plays the advertisement file. The playback mode is generally based on the auxiliary screen, and does not affect the streaming content that the user is watching. The location where the advertisement information is played on the screen can be arbitrary or active, such as a drift advertisement.

4 is a schematic structural diagram of advertisement classification storage information 210. Index 410 is a serial number or ID number of the advertisement classification storage information, and can be used to identify different advertisement information items. The classification information 420 is a method for classifying and storing the advertisement information. By this method, any kind of advertisement object can be attributed to a complete classification information. The advantage is that the location or the location of the specific advertisement object in the advertisement classification storage information 230 can be found by the method of hierarchical indexing, and the advertisement information of the similar advertisement object can also be found in the same level or the primary index. Using this hierarchical indexing method, similar advertising information can be pushed to the user. For example, Coca-Cola's four-level index is carbonic acid, so it can be found that the four-level index is carbonated and Pepsi-Cola, and Pepsi and Coca-Cola have strong correlation. This correlation can be used when pushing ads to users in the future, and pushing the same or similar content-related products is equally easy for users to receive.

In FIG. 4, the search information 430 is the focus of the patent, and the video, audio, and character information that can be extracted from the streaming media file are equally divided into three types of images 432, 431, and 433 that can be corresponding. The information recorded by the image 432 includes the trademark of the advertisement, the shape or important feature or logo of the advertisement product, some image information related to the advertisement product, and the image associated with the advertisement product, if the origin of the advertisement product or the related person image, etc. The search information is recorded in the image 432-column. The same record for the sound includes the pronunciation of the name of the advertising product, the pronunciation of the advertisement, and the pronunciation of the place name or the name of the person associated with the advertisement, and any sound recording associated with the advertisement product. Search information 430 Chinese text 433 typically has some relevance to sound 431, and the record of text 433 includes any product-related text such as product name, advertising word, related person name or place name. This correlation between sound and text is also reflected when matching speech or characters in streaming media content. For example, if the text is a subtitle form of a streaming media file, then the text and sound have a high correlation. Searching for information 430 usually requires careful processing of the advertising information, and it is possible to add a lot of artificially collected information to complete the lack of search information. The search information will provide as many similar relationships as possible for each ad item. The more complete this similar relationship, the easier it will be to find a matching location in the streaming content, and more information can be inserted. 440 is the location/attribute column, which includes two sub-sections of link and content. The link sub-column provides location information for extracting advertisement information. For example, Coca-Cola needs to be obtained through a specific HTTP address, and the location of such link can be remotely located. Or just local to the user's machine. The content sub-bar is mainly to indicate the attributes of the ad item itself, such as video ads, audio ads, image ads, text ads, and the like. With the advertisement classification storage information, it is possible to provide the content analysis module 120 with the possibility to find accurate advertisement information for the content. Here, the image, the sound, and the text are the iconic search information, but some of the advertisement files only include one or two of the image, the sound, and the text. For example, an advertisement contains only one image, that is, image information. This will result in insufficient search information. In order to solve this problem, it is necessary to supplement the search information to achieve the image, sound, and text information as much as possible, and increase the capacity of the matching information, which can greatly improve the search. The hit rate of the content of the streaming media file matches the content of the advertisement. For example, in a Coca-Cola advertisement containing only the promotional image of Coca-Cola, the Coca-Cola pronunciation and the "Coca-Cola" text can be added to the search information as the iconic search information. In many cases, if the content of the search information is insufficient, it is often necessary to add sound, image, and text information related to the advertisement content other than the advertisement information. The added content may be from the website of the advertisement product, and the sound may also be manually recorded. Sound or from a web clip.

FIG. 5 is a schematic diagram of the content analysis module. The three streams of the video stream 510, the audio stream 520, the character stream 530, and the 4 bar are extracted from the streaming media file 220 and input into the content analysis module 120. The video stream 510 and the image 432 column information of the search information 430 and the location/attribute 440 in the advertisement classification storage information 210 simultaneously enter the video analysis module 540, where all the image column information in the search information column in the advertisement classification storage information 210 is input. That is, all possible image information related to the advertisement is input to the video analysis module 540, and each image 432 record is associated with a location/attribute 440 record of an advertisement. It should be noted here that in the actual application, each sub-column of the search information 430 column, such as the image sub-column, and the position/attribute record correspondence may be many-to-one, one-to-one, one-to-many, three possible . The video analysis module extracts each frame in the video stream 510 and analyzes whether there is a match or partial match with the information recorded in the image 432 column of the search information 430 in the advertisement classification storage information 210, if any, corresponding to the image advertisement. The location/attribute 440 generates a portion of the information code to be added to the advertisement description file 230. The added information code can be in the following format:

[Location or area value of the frame in the video: Time: Link: Ad content format: Open Location]

The position or area value of the frame is to indicate the position of the position where the advertisement is added in the entire video, such as the area value of the first few frames or a certain frame. The time parameter can be the start to the knot Bunch time. The link is the address information stored in the advertisement. The format of the advertisement content refers to whether the advertisement is an image or a sound or other format, which is obtained in the position/attribute 440 column. The open position is based on the ad properties (obtained in the location/attribute) to set the open position and open mode, if you open the Media Player software in the upper right corner, etc.

This format is also the primary means of composing an ad description file.

In the same manner, audio stream 520 is input to the audio analysis module along with the location/attribute 440 associated with all sound 431 recordings and sound recordings. If a portion identical or similar to the sound 431 record is found in the audio stream, a portion of the information code is generated to be added to the advertisement description file 230. The information code generation process and the process of generating the information code by the information code format and the video analysis module are similar to the information code format.

Similarly, character stream 530 is entered into the character analysis module along with the location/attribute 440 associated with all text 433 records and text records. If a portion identical or similar to the text 433 record is found in the character stream, a portion of the information code is generated to be added to the advertisement description file 230. The information code generation process and the information code format are similar to the process in which the video and audio analysis modules generate the information code. The information code format is similar, as follows:

The information code is only an important part of the advertisement description file, and the advertisement description file can have a standard markup language format, similar to XML or XSL. The information code expression format is just an abstract textual expression that summarizes the main elements of the information code. These elements can be appropriately added or deleted. For example, the position and time of the frame often have an intrinsic correspondence. Frame retrieval of the ad and retrieval of the ad at a fixed time can achieve the same effect, so only one of the frame position or time can work. It is also possible to add new elements such as advertiser information, player information, encryption or DRM information, and more. The order of the elements in the information code is also not fixed, and the order of the elements can be adjusted as needed. The language used for the information code can use the class XML (Extensible Label Language) or XSL (Extensible Style Language). The information code can also be in encrypted or binary format.

There are several modes for analyzing video and audio in the video and audio analysis module: The first type is the manual recognition mode. At present, image recognition technology can not identify any object in general, and intelligently use modeling to identify some special objects, for example, face recognition. Therefore, the introduction of manual identification of advertising objects at this stage is undoubtedly a relatively simple method. When manually identifying an object, it is necessary to manually check whether there is image content matching or similar to the advertisement information in each frame. For example, it is known that there is Coca-Cola in the advertisement information, when the staff finds a Coca-Cola logo or product at a certain position of the video, It is then possible to insert the location/attribute 440 information of the advertisement at this location and generate an advertisement description file 230 in conjunction with the location or time information. The location can be a specific frame or a specific location within the frame. At this time, the information of the image 432 column in the search information can be omitted or used as a reference for manual inspection.

The method of analyzing audio using the manual recognition mode is similar to video, except that there is a difference between whether human hearing or visual use is used to distinguish whether or not there is advertising information.

Second, automatic identification. For the recognition of sound, the technology is now mature. As long as the speech in the search information 430 is extracted and then the recognition training is performed, a high recognition accuracy can be obtained. The commercially available products for automatic voice recognition include VoiceTone, Nuance OSR, ViaVoice, e-talk, etc. These products all have strong voice recognition functions, and only need to combine these products with the voice analysis module to accurately implement voice. Identification function. Some existing commercial products can also be used for image recognition, such as Excalibur Technologies, Mediasite and Virage, and the like. For video, only the other processing is similar to the audio recognition process. In the case of automatic identification, there is often a problem of similarity judgment, usually by setting similar parameters to solve this problem.

When the character analysis module performs character stream analysis, the information of the text 433 column in the search information can be accurately matched or similarly matched, and the position where the matching is successful plus the position/attribute 440 is used to generate the information code. Description file. In some cases, the audio stream 520 and the character stream 530 have a high correlation. For example, the character stream is highly matched for the subtitle file and the audio stream. At this time, only the character stream needs to be analyzed, and the information code is generated by adding the advertisement description file. It is.

As shown in FIG. 6, the flow inside the user playback device 190 is shown. The user playback device decodes the streaming media file through the streaming media decoding module 610, and after decoding The video is presented to the user through the video window. At the same time, the ad description file parsing module

620 parses the received advertisement description file 230, and the advertisement operation module 630 mainly performs the function described in the advertisement description file, for example, at a certain time or a certain frame position, retrieves the advertisement at an advertisement storage server, and is in the user The ad is played at a location on the playback window of the playback device. The window for playing the advertisement may be a window for playing the streaming media file, which is to pause the playback of the streaming media file; or in a multi-window environment, the streaming media playback and the advertisement playback each have a playback window; An ad window is used to play the ad. The advertisement file 640 is typically stored in an advertisement storage server that retrieves the desired advertisement file from the advertisement storage server when the user desires an advertisement.

As shown in FIG. 7, it is a schematic diagram of the internal process of the advertisement storage server 140. The advertisement file 640 will typically be a database storage system, and each advertisement file will then be extracted for placement in the advertisement processing module 710, and finally the advertisement classification storage information 210 will be generated by the advertisement processing module. There is also an advertisement extraction interface 720 in the advertisement storage server, and the user playback device communicates with the advertisement extraction interface, extracts the advertisement file, and plays on the user playback device.

FIG. 8 is a schematic diagram of a process of expanding advertisement search information in an advertisement storage server. The information found in the extended advertisement is mainly completed in the advertisement processing module 710. In fact, this is a new search method. The focus of this search method is not only to find the object being searched, but to find out the information related to the object being searched. The system is actually just one example of the new search method, including the following steps:

810. The advertisement description file is first extracted by the advertisement processing module 710.

820, determining whether the advertisement information is sufficient, if it is sufficient, recording in the search information 430 in the advertisement classification storage information 210; if not, proceeding to step 830;

830, increase the ad search information, and then continue to determine whether the ad information is sufficient. Finally, the advertisement classification storage information is sent to the advertisement analysis module. In the process of judging whether the advertisement information is sufficient, it is usually a one-to-one judgment, for example, first judging the video or image information, judging the audio information, and finally judging the text information. When you add advertising search information, you can also use this one-to-one method to add categories. Of course, you can also add together to judge. The way to increase advertising search information is mainly to add as many missing key points as possible. Information, the key information here includes: images, sounds and text. The reason is that any advertising information can be summarized in one or more of these three. The image contains continuous images, that is, video, and also contains any image related content such as flash animations and icons.

9 is a schematic diagram of a user play device play window, which includes a main window 910 and a sub-window 920. The embodiment window 920 includes three sub-windows A, B, and C. The main window 910 is usually a streaming media file 220, and the sub-window. 920 is typically played with an advertisement file 640. The characters in the main window 910 are wearing shirts 930, shorts 940 and shoes 950, corresponding to the advertisement of the shirt, the advertisement of the shorts, and the advertisement of the shoes in the advertisement window. For example, when a streaming media file is played to a picture frame in the main window 910, the sub-window 920 automatically plays an advertisement for the shirt, shorts, and shoes associated with the content in the main window 910. 9 is only a conceptual design. In a specific implementation, the positional relationship between the sub-window 920 and the main window 910 can be very flexible, for example, the sub-window is in the main window, or the sub-window is on one side of the main window. The child window can also be switched to the main window if needed. These are all prior art and are not specifically described herein.

FIG. 10 is a schematic diagram of a user request index, which usually includes information such as user 1010, advertisement object 1020, number 1030, and priority 1040. First, you need to generate a user request index, which is to create a click or watch advertisement record for each user. In Figure 10, Coca-Cola is clicked or viewed 7 times. After the user requests the index to be established, the user can push the advertisement service according to the user request index and the advertisement classification index.

There are two ways to push ads:

One is user-centric, pushing ads to fixed users. When the advertisement is selected for the A user, all the clicked or viewed advertisements of the A user are retrieved, and the highest priority or the second highest advertising target is selected as the benchmark advertisement, and then searched in the advertisement classification index. The location of the benchmark advertisement (such as Coca-Cola), after finding the characteristics of the advertisement classification index, select the relevant advertisement content, and then push the relevant advertisement content to the A user.

The other is advertising-centric, pushing an ad to the most acceptable user of the ad. For example, get a Coca Cola ad that needs to be pushed to potential users. First find the relevant ads in the ad category index and the ads to be pushed, as shown in Figure 4. Find Coca-Cola, Pepsi, etc.; then go to the user request index to find and push the relevant ads of the advertisement, and then find the user corresponding to the relevant advertisement according to the priority from high to the bottom, called the target user; Finally, put the advertisement to be pushed For the target user.

The calculation method of the priority P is usually: where P represents the priority, and represents the coefficient to be determined, and the undetermined coefficient may be 0, X represents the number of clicks or views by the user, that is, the number of times of interest, and Y represents the broadcast of the advertisement itself. The value of the advertisement, such as the value added, Z, the cost of playing the advertisement, such as the length of the advertisement. The relevance of the advertisement classification index is that in the classification information of Figure 4, the correlation from the four-level online network is decreasing, which is determined according to the classification method, that is, the classification is continuously reduced in the first-level to the fourth-level classification. region.

Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware or by software plus necessary general hardware platform. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several The instructions are for causing a computer device (which may be a personal computer, server, or network device, etc.) to perform the methods described in various embodiments of the present invention.

In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Claims

Rights request

A method for adding advertising information to a streaming media, comprising the steps of:

Opening and processing a streaming media file, adding advertising information related to the content of the streaming media file to the streaming media, and generating an advertisement description file;

Sending the streaming media file and the advertisement description file to the user;

The user playing device plays the streaming media file, and parses the advertisement description file corresponding to the streaming media file;

The user playback device retrieves the advertisement file and plays the advertisement according to the advertisement description file.

2. The method according to claim 1, wherein the adding the advertisement information related to the content of the streaming media file in the streaming media, and generating the advertisement description file comprises the following steps:

Extracting video and/or audio and/or character content from the streaming media file, determining whether the content matches the search information in the advertisement classification storage information, and if so, generating an information code and adding an advertisement description file;

If there is no match, continue to search and determine if the remaining video and / or audio and / or character content matches the search information in the ad classification storage information;

The content correlation includes: one or more correlations of image correlation, sound correlation, and text correlation.

The method according to claim 1, wherein the advertisement description file includes one or more of a location where the advertisement file is stored, a location where the advertisement file is played, and a time when the advertisement file is played;

The location where the advertisement file is played refers to playing an advertisement file when the streaming media file is played to a specific certain frame.

The method according to claim 1 or 2, wherein at least one of image recognition, voice recognition, and character recognition is required for the streaming media file when determining the correlation between the streaming media content and the advertisement content.

5. A system for adding advertisements to streaming media, characterized in that: a content analysis module, configured to analyze the content of the streaming media file and the relevance of the advertisement content, including a database for storing the advertisement classification storage information, and connected to the storage streaming media file server and the advertisement storage server;

An advertisement storage server for storing advertisement information and connecting with a playback device and a content analysis module;

a playback device, configured to parse the advertisement description file at the user end and extract the advertisement description file according to the advertisement description file, and the playback device transmits the advertisement storage server and the storage streaming media file and the server or module storing the advertisement description file through the network Connected

A streaming media storage server for storing streaming media files and connecting to a content analysis module.

The method according to claim 5, wherein the advertisement description file is stored in the same server together with the corresponding streaming media file, or the advertisement description file and the corresponding streaming media file are respectively stored in different servers.

A method for generating an advertisement description file, comprising the steps of: generating the advertisement item search information according to each of the stored advertisement content items, the search information including sound, image, and text;

Generating location information of the advertisement item based on each of the stored advertisement content items;

The search information is used to determine whether the content of the streaming media file matches the stored content of the advertisement. If a matching relationship is found, the location information of the advertisement and the location or time information of the matching of the streaming media are generated into an advertisement description file.

8. The method according to claim 7, wherein the client obtains advertisement information related to the streaming media content by parsing the advertisement description file.

9. The method of claim 7, the advertisement description file comprising an XML-like format or a XSL-formatted scripting language.

10. A method for increasing file relevance search hit ratio, characterized by comprising the steps of:

Processing the search file to obtain the iconic search information of the search file; The iconic search information of the search file is expanded by the correlation processing; the searched file is searched by the iconic search information, and whether the searched file has information matching the iconic search information is determined.

11. The method according to claim 10, wherein the entropy search information of the search file is expanded by correlation processing, comprising the following steps:

If the iconic search information extracted directly from the search file contains only one or two of the sound, the image, and the text, the missing sound and/or image and/or text information is added to the search information.

The method according to claim 11, wherein the adding missing sound and/or image and/or text information includes sound, image, and text information from outside the file.

13. A method for pushing relevant advertising information to a user centered on a user, comprising the steps of:

Finding the user who needs to push the advertisement information in the user request index, and selecting the standard advertisement to be pushed according to the priority of the advertisement of the user for all the users, and searching for the relevance in the advertisement classification index according to the benchmark advertisement content. High related advertising.

Advertise relevant ads to users and play them on the client side.

14. A method of pushing advertisement information to potential users centering on advertisements, comprising the steps of:

According to the advertisement information to be pushed, find the relevant advertisement in the advertisement classification index; then find the user who is concerned about the related advertisement in the user request index, and according to the priority of the user concerned about the related advertisement And select a user with a higher priority as the target user of the push advertisement;

Push the ad to be pushed to the target user and play it on the client.