WO2007128234A1 - Procédé et noeud de filtrage de flux vidéo - Google Patents

Procédé et noeud de filtrage de flux vidéo Download PDF

Info

Publication number
WO2007128234A1
WO2007128234A1 PCT/CN2007/001463 CN2007001463W WO2007128234A1 WO 2007128234 A1 WO2007128234 A1 WO 2007128234A1 CN 2007001463 W CN2007001463 W CN 2007001463W WO 2007128234 A1 WO2007128234 A1 WO 2007128234A1
Authority
WO
WIPO (PCT)
Prior art keywords
module
harmful
content
filtering
intra
Prior art date
Application number
PCT/CN2007/001463
Other languages
English (en)
Chinese (zh)
Inventor
Zhong Luo
Original Assignee
Huawei Technologies Co., Ltd.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co., Ltd. filed Critical Huawei Technologies Co., Ltd.
Publication of WO2007128234A1 publication Critical patent/WO2007128234A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/16—Analogue secrecy systems; Analogue subscription systems
    • H04N7/162—Authorising the user terminal, e.g. by paying; Registering the use of a subscription channel, e.g. billing
    • H04N7/163—Authorising the user terminal, e.g. by paying; Registering the use of a subscription channel, e.g. billing by receiver means only
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/442—Monitoring of processes or resources, e.g. detecting the failure of a recording device, monitoring the downstream bandwidth, the number of times a movie has been viewed, the storage space available from the internal hard disk
    • H04N21/44209—Monitoring of downstream path of the transmission network originating from a server, e.g. bandwidth variations of a wireless network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/454—Content or additional data filtering, e.g. blocking advertisements
    • H04N21/4542—Blocking scenes or portions of the received content, e.g. censoring scenes

Definitions

  • the present invention relates to multimedia communication technologies, and in particular, to a video stream filtering method and a filtering node in a multimedia communication process. Background technique
  • Streaming Media As a basic form of multimedia communication, Streaming Media has spawned many forms of multimedia communication services: conference TV/videophone, IPTV, VOD, instant messaging and more. Therefore, streaming media will become the basic form of communication on the Next Generation Network (NGN).
  • NTN Next Generation Network
  • IPTV Internet Protocol Television
  • IPTV and VOD Video on Demand, video-on-demand
  • VOD Video on Demand, video-on-demand
  • the content is very broad, including film and television programs, news, sports competitions, concerts and more.
  • operators/ISPs Internet Service Providers
  • the domestic IPTV operation will be launched on a large scale. The first question is how to ensure effective content monitoring and filtering, and to shield harmful content from the problem.
  • the IPTV operation in China will not be discussed, and the relevant national departments will not A license may be issued. Therefore, the solution of this problem is of great significance to promote the development of the IPTV industry.
  • the usual understanding includes two aspects:
  • the object of protection is the object of content attack. Usually an audience. '
  • Content filtering is the processing and judgment of certain attributes of content. These content attributes can include: the name of the content provider, the URL of the content (Universal Resource Locator, the URL is an important type of URL), content The IP address of the server, etc., and the packet header information of the packet in the case of the packet encapsulation in the case of the packet, the information in the packet, and the like. It can be seen that this processing and filtering is also carried out in a hierarchy from shallow to deep.
  • the prior art one mainly performs content filtering according to external features of the content, or shallow features.
  • the most typical example is URL-based filtering.
  • the principle is shown in Figure 1:
  • the content filtering device is located between the core network and the edge access network on the network, so that the media stream from the content source arrives between the receiving terminals.
  • NAT Network Address Translator
  • FW Firewall
  • BAS Broadband Administration System
  • BRAS Broadband Registration and Admission System
  • DISLAM Broadband Registration and Admission System
  • ISP's POP Point of Presence
  • the filtering device itself has an internal database, and there are multiple content source URLs, according to this number.
  • the database can determine whether a part of the content source is harmful, and block harmful content sources and release harmless content sources.
  • content tiering service providers that provide third-party services. Their databases are more abundant and professional.
  • Content filtering devices can also connect with such third-party service providers to use their services for U L-based filtering.
  • Some URLs may be considered as qualified websites in the grading system, and may also have problems (being hacked to impersonate their websites, or their own illegal attempts, etc.);
  • DPF Deep Packet Filtering
  • the prior art 2 deep DPF is based on manual deep content setting.
  • the content filtering device can decode the media stream and play the content (assuming that encryption is not a problem, because the encryption problem can be legally monitored by the communication device. Request for resolution), for review by manual monitors. If a problem is found, the monitor immediately takes action to cut off the harmful content and switch to a harmless content such as a public service advertisement.
  • the content filtering device there must be a relatively large capacity delay device to delay the harmful content, giving the monitoring personnel certain judgment and reaction processing time (for example, 5 seconds). .
  • the prior art 2 has the following problems:
  • the invention provides a video code stream filtering method and a filtering node in a multimedia communication process, so as to solve the problem that the existing manual content-based deep content filtering method has low efficiency and lacks universality.
  • the video code stream filtering method in the multimedia communication process of the present invention includes the following steps:
  • the video code stream filtering node in the multimedia communication process includes: providing a video code stream in a multimedia communication process
  • the filtering method only needs to partially decode the I frame image in the video stream or the I frame and the image of a certain number of frames adjacent to and/or after the decoding, and does not need to decode the image of most other frames, which reduces the complexity of processing. Degree, shortening the delay time of video stream playback, and improving the efficiency of video content depth filtering;
  • the present invention is further based on a scene segmentation technique, partially decoding a first frame image in each scene or an image of the first frame and a certain number of frames adjacent to and/or after the previous frame, and using the decoded image for recognition, in ensuring recognition Reliability also reduces the number of data frames that need to be decoded to a certain extent, so that the processing complexity is further reduced;
  • the method of the invention is based on the automatic identification technology of the existing harmful content, can realize the automatic identification and filtering efficiently, and ensures the rapid and effective identification of common harmful content;
  • the method of the invention can be used in combination with the artificial identification mechanism at the same time, which can prevent the missed detection of new harmful internal winter;
  • the invention also provides a harmful content learning mechanism, which can be learned and added to the harmful content database when manually identifying new harmful content;
  • Most of the methods of the present invention can simultaneously adopt the existing UTRL-based filtering technology, and can prohibit the source of harmful content at the signaling level. Moreover, the invention further provides a URL information rating mechanism for harmful content, and can gradually discover new harmful URL sources. , and add new harmful URL sources to the harmful URL repository in a timely manner;
  • the method of the present invention also provides a log reporting mechanism, which can record various events in the video stream filtering process;
  • the video code stream filtering node of the present invention can conveniently implement the method of the present invention, and has Very good versatility;
  • the technical solution of the present invention can solve the content security problem in the multimedia services such as IFTV and digital television, and ensure the security and reliability of these services.
  • FIG. 1 is a schematic diagram showing the principle of filtering based on a content-based URL
  • FIG. 2 is a schematic diagram of a relationship between a frame and a scene in a video sequence according to the present invention
  • FIG. 3 is a schematic diagram of correspondence between data packets in a scene, a frame, and a video code stream according to the present invention
  • FIG. 4 is an exemplary diagram of a feature network model according to the present invention
  • FIG. 5 is a schematic flowchart of a video stream content filtering method according to the present invention
  • FIG. 6 is a schematic diagram of a main structure of a video stream filtering node for implementing the video stream filtering method of the present invention. detailed description
  • the invention provides a video code stream filtering node (Node) which is arranged at a suitable position in the network, and the filtering node can implement automatic filtering and manual filtering on the content in the streaming video, and can simultaneously filter based on URL or the like.
  • the shallow filtration method is used for filtration.
  • the automatic filtering method of the video content of the present invention is first given.
  • the automatic filtering method of the present invention uses the I frame in the video code stream as the object to be detected, and restores the I frame image after decoding the I frame to identify the harmful content.
  • the I-frame identifier is set in the header of the packet containing the I-frame, which can be identified.
  • FIG. 2 is a schematic diagram showing the relationship between a frame and a scene in a video sequence.
  • scenes For a video stream that passes through a filtering node, it is first divided into different scenes (Scene), which is originally used as a
  • a video sequence consisting of multiple frames is divided into sequence sequences of different scenes.
  • a scene contains a number of frames, and each frame inside each scene is basically the same in the background and foreground, but there is a certain motion.
  • For split scenes it must be stated that the scene was originally created when video content capture (lens switching) and production (adding effects such as 3D transition effects between two shots).
  • Performing scene segmentation on the filtering node is to divide the code stream in the video stream into segments, each segment corresponding to the original scene.
  • the scene that is finally split on the filtering node may not be completely consistent with the scene inherent in the video stream, but does not affect the application of the present invention.
  • FIG. 3 is a schematic diagram of correspondence between data packets in a scene, a frame, and a video stream, because the video stream is sent from a device such as a streaming media server (Streaming Media Server), and is compressed. Packetization, regardless of the specific packaging protocol, is sent in chronological order. Each packet has a corresponding serial number or timestamp (Time Stamp, etc.). Based on this information, the node can be correctly weighted. The original order of the composition, so that the package corresponds to the scene. Therefore, the final result is a scene corresponding to a series of video packets.
  • Streaming Media Server streaming media server
  • the filtering node only needs to identify the first frame of each scene, so that all the scenes can be segmented, and all the frames between the first frame of one scene and the first frame of the next scene belong to the scene.
  • I frame intra coded frame
  • the so-called I frame is for a P (predictive coded frame) frame and a B frame (bidirectional predictive coded frame).
  • P predictive coded frame
  • B frame bidirectional predictive coded frame
  • I frames are added when the scene changes, and the first frame of the scene is often an I frame.
  • a new standard such as H.264
  • a complete I may not exist
  • some modified selection criteria can be defined: for example, selecting a frame with an intra-coded slice or a macroblock MB (Macroblock).
  • Macroblock macroblock MB
  • identification mechanisms to identify I-frames or intra-coded stripes.
  • the filtering node can correctly extract the I frame or the intra-coded strip/macroblock according to these specific identifiers.
  • adjacent images of adjacent frames before and after the I frame can be partially decoded at the same time, which are used to assist in identifying the I frame image. According to experience, in most cases, 5 frames can be accurately obtained. The purpose of identification is. Of course, when the accuracy of the decoded I-frame image cannot be accurately identified, the adjacent images of adjacent frames before and after the I-frame are partially decoded to assist in identifying the ijl frame image.
  • I frame For convenience of description, the following describes an I frame as an example. It may be in one scene (the lens is long), there are multiple I frames, then specify the first I frame in a scene.
  • the filtering node After obtaining the first I frame in a scene, the filtering node decodes the I frame and restores the I frame image, and then identifies the frame image, including the following two identification methods:
  • the automatically identified harmful content includes the following:
  • the (Optical Character Recognition) module recognizes.
  • the recognition result is matched with the database, and if the harmful judgment condition in the database is successfully matched, it is determined to be a harmful superimposed text or symbol, and the superimposed text or symbol recognition technology belongs to a mature prior art;
  • the recognition result of the automatic identification module shall be taken as the standard.
  • One embodiment is: A weighted average based on scores. Automatic identification modules and human monitors not only determine whether they are harmful, but also give harmful scores, such as from 0-100. The higher the degree of damage, the higher the score, and 0 means harmless. Then add the score of the automatic identification module and the score of the human monitor to the following:
  • W M and W H represent the weights of the automatic identification module and the human monitor.
  • the relative size between the two indicates that the automatic recognition module is still more human or human, and S M and S H respectively represent the scores given by the automatic recognition module and humans. If the resulting composite score is greater than a given value, such as 50, then the joint judgment is harmful, otherwise it is harmless. If only one party identifies the harmful content and gives the score of the harmful content, you can default the other party. The score given for this content is 0.
  • the action taken can be:
  • the filtering node should also have a learning function. If the harmful content is not automatically discovered by the automatic identification module, but is discovered by the human monitor or discovered through other channels, then the learning module in the filtering node must learn to write the harmful video stream. . In order to learn the system, each monitored code stream needs to be stored for a certain length of time (for example, 10 minutes, considering the required capacity, this time length should be optimally adjusted). In order to further reduce the required storage capacity, only I frames for identification can be stored for each scene. Once a human monitor finds harmful content happening at a certain moment Before and after, the learning module should read the I frame of the corresponding scene from t-TW/2 to t+rw/2 (the length of the learning time window, for example, 30 seconds) from the database for learning. After learning, the automatic identification module can identify such related scenes later. There are many ways to learn, including Artificial Intelligence, Fuzzy Logic, and Artificial Neural Network.
  • the filtering node When filtering based on URLs at the same time, the filtering node also "remembers" the harmful content from the content source and other related information, stores it in the corresponding "suspect” database, and ranks the URL and other related information according to the history. For URLs stored in this "suspect" database, some more granular processing is required. If a legitimate URL is only because of some mistakes or is faked by someone else to play the harmful content, then if it is stored in the "suspicious" database, as long as it does not happen in the future, after a period of time can eliminate its "suspect", instead, If you find the bad behavior of a URL multiple times, you can determine it as "blacklist," and thus completely shield it. You can also share the information with the database of the third-party URL rating service provider, and send the identification result of the filtering node. Give the third party a rating service provider database so that you can work together for mutual benefit.
  • the scene segmentation technology used in the present invention generally includes the following two types:
  • a two-level video harmful content filtering technology can solve such problems.
  • the basic idea is to hierarchically define image features, which are generally divided into two large layers, namely, semantic (Semantic or Conceptual) level features, and event level features.
  • semantic Semantic or Conceptual
  • event level features For example, as shown in FIG. 4, if the highest semantic feature to be detected is an "outdoor scene", the corresponding lower-level semantic features include “beach”, “mountain forest”, “wild field”, etc., and further corresponding to lower levels. Semantic features, and finally to event features, such as a mountain, or a piece of trees.
  • Each event feature has specific identification methods, such as identifying roads, people's movements, and so on.
  • the advantage of using this two-layer identification method is that it combines low-level features that can be automatically identified with advanced features that humans can understand. Such correspondences can form a feature network model.
  • a feature network model of concepts such as “pornography” and “violence” can be established.
  • the establishment of a feature network model needs to be based on human understanding of the cognitive process mechanism and expert knowledge of a specific field. It belongs to the prior art, and the present invention does not further description.
  • the present invention provides an input interface through which a human expert can define an expression of a feature network model, and the filtering node can perform automatic identification based on the feature network model.
  • the filtering method for the harmful superimposed characters and graphic symbols used in the present invention can locate the superimposed subtitles and graphic symbol regions in the image without decoding, and then extract them, and after a certain background foreground segmentation, input an OCR (Optical Character Recognition, Optical Character Recognition)
  • OCR Optical Character Recognition, Optical Character Recognition
  • the module recognizes the discrete Cosine Transform (DCT) coefficients in the data stream in the video stream, and can determine the rectangular area containing the superimposed text or graphic symbols.
  • DCT discrete Cosine Transform
  • For the horizontal and vertical projection of the area Projection, in fact, all the horizontal or vertical lines passing through the area, the integral of the pixel brightness on the line is summed to obtain a one-dimensional brightness distribution curve), judge text or The direction of the symbol, and then the use of a similar projection method for the division of lines and words.
  • the filtering node of the present invention can also implement a log recording function and is connected with an external control device to implement data and signaling interaction with the external control device.
  • the present invention first provides an I-frame based deep content filtering method, as shown in FIG.
  • the identification processing method for each I frame to be detected includes the following steps:
  • the I frame to be detected may include each I frame in the video code stream, and is identified according to a correspondingly set I frame identification signal in a packet header of the data packet including the I frame;
  • the first frame in a scene is generally the I frame of the scene.
  • the I frame refers to the frame containing the intraframe encoded strip or the macroblock MB, and the frame identifier has an instant. Decode the refresh IDR flag.
  • the replacement video source can also be started to be played.
  • the video stream is first segmented before the frame to be detected is acquired, and then the first frame in each scene is used as the frame to be detected, and the first frame or the frame is partially decoded. And a certain number of frames that are adjacent to and/or after.
  • the method of the present invention can be used in conjunction with existing URL-based filtering.
  • URL-based filtering can filter related signaling in a multimedia communication process. If the relevant signaling contains harmful URL information, the signaling is refused. , thereby preventing the reception of video streams from harmful URL sources.
  • the invention provides URL-based filtering, and also provides a harmful URL information rating mechanism, which can prevent accidental killing of harmful URL information, discover new harmful URL information, and then add newly discovered harmful URL information to harmful URL information in time.
  • a harmful URL information rating mechanism which can prevent accidental killing of harmful URL information, discover new harmful URL information, and then add newly discovered harmful URL information to harmful URL information in time.
  • the specific identification method can adopt manual identification and automatic identification, generally
  • the two identification methods can be used to obtain higher efficiency and more secure identification.
  • the judgment result of performing manual identification or automatic recognition can be set preferentially.
  • the judgment results of both parties can also be comprehensively considered to obtain A more responsible control mode.
  • the method of the present invention also provides a content recording mechanism, including recording of the identified harmful content and recording of the video stream played during the specified time period, and recording the harmful content is: if there is no harmful content stored in the automatic identification
  • the information is manually identified, and the present invention also provides a learning mechanism to ensure that newly generated harmful content is added to the harmful content database of the automatic identification mechanism in time; the video stream recorded during the specified time period is recorded as follows: specific
  • the video stream of the U L source, or the harmful content in the video stream, is missed and provides information for future learning.
  • the method of the present invention also provides a logging and reporting mechanism to record the filtering process of the video stream and generate a log report.
  • the video stream filtering node provided by the present invention mainly includes:
  • a video stream delay module configured to receive a video code stream to be played during a multimedia communication process and delay outputting the video code; the specific delay time is determined according to a required experience time for identifying the harmful content; and a switch module, connecting the video code stream a delay module, configured to cut off a video code stream output by the video stream delay module;
  • An I frame detection/decoding module is configured to obtain, from the video stream to be played in the multimedia communication process, adjacent frames before and after the I frame I frame to be detected, and partially decode the I frame image to be detected and before the I frame And adjacent adjacent frames of adjacent images;
  • the adjacent frames before and after the I frame are obtained from the video stream delay module, and the phase before and after the I frame is partially decoded. Adjacent frames of adjacent images are used to assist in identifying the I frame image.
  • the I frame detection/decoding module is simultaneously connected to the video stream delay module.
  • a harmful content identification module connected to the I frame detection/decoding module, configured to identify whether the I frame image contains harmful content, and if yes, output a corresponding control signal
  • a decision module connected between the harmful content identification module and the switch module, configured to output a trigger signal for disconnecting the video code stream to the switch module when receiving the control signal
  • the filtering node further includes: a scenario segmentation module, configured to connect the I-frame detection/decoding module, and receive the video stream to be played in parallel with the video stream delay module and the video The code stream performs scene segmentation;
  • FIG. 7 is a schematic diagram of a mechanism of a harmful content identification module and a decision module, wherein the harmful content identification module includes:
  • An automatic identification sub-module implementing an automatic identification function is connected between the I-frame detection/decoding module and the decision module for comparing the harmful content in the harmful content database with the related content included in the I-frame image. Automatic identification of harmful content for the purpose;
  • the automatic identification sub-module further includes: a harmful image recognition unit and a connected harmful image database, a harmful superimposed text/symbol recognition unit, and a connected harmful superimposed text/symbol database, a face recognition unit, and a connection according to the type of the harmful content.
  • the face database in parallel, identifies whether the corresponding harmful content is included in the I frame image.
  • the harmful image database also stores various existing feature networks for identifying harmful image content (each time the feature network input by humans is stored here) and various templates for identifying low-level event features, such as statistical histogram templates;
  • the text and symbol database stores templates for various harmful words and symbols, such as reactionary and erotic vocabulary slang, as well as known harmful graphical symbols, such as Vietnamese symbols;
  • the face database provides the necessary data and various types for the face recognition module.
  • the harmful content identification module further includes a manual identification sub-module for implementing the manual identification function
  • the artificial identification sub-module specifically includes: an I-frame image display unit and a monitoring instruction input unit, wherein the I frame image display unit is connected to the I frame detection/decoding module, configured to display the I frame image to a monitor for manual identification of harmful content; and the monitoring instruction input unit is connected to the decision module , used to receive the monitor to identify harmful When receiving the cutoff command when the input to the decision module outputs the control signal.
  • the corresponding in the decision module includes:
  • a first determining unit receiving a control signal output by the automatic identification submodule
  • a second determining unit receiving a control signal output by the operation interface submodule
  • a joint determining unit configured to respectively connect the first determining unit and the second determining unit, configured to preferentially execute a control signal of the first determining unit or the second determining unit according to the set rule; or, the automatic identifying submodule and the monitoring
  • the corresponding harmful degree scores are given for the identified harmful content
  • the joint judgment unit weights the two scores to obtain the final executed judgment result, and only one party is recognized.
  • the harmful content gives a score
  • the default score given by the other party for the content is zero;
  • the module can introduce the audio content filtering result from the outside of the node as an input, so the decision module further includes:
  • a third determining unit configured to output a trigger signal for disconnecting the video code stream to the switch module directly when receiving a harmful sound decision result of the audio code stream corresponding to the video code stream, or by using a joint decision unit
  • the switch module outputs a trigger signal for disconnecting the video code stream, and the structure shown in FIG. 7 is the latter implementation.
  • the video stream filtering node can also include:
  • a URL-based filtering module configured to receive related signaling of the multimedia communication, and perform URL-based filtering on the related signaling by using a pre-stored harmful universal resource locator URL information library, and if it is determined that a certain URL is harmful, prohibiting Corresponding signaling establishment process, so that the request and transmission of content cannot be performed correctly;
  • the filtering node may further include: a URL recording and rating module and a URL rating database, wherein the URL recording and rating module is configured to record URL information of harmful content, the URL rating database is used to record URL rating data; and the URL recording and rating module is based on a The frequency and severity of the bad behavior of the URL before, the rating adjustment is made, and if the URL information reaches the set level, the URL information is added to the harmful UL information base. This guarantees that it will not be blocked forever due to a problem with a URL, and that it can also output records and rating results to third-party rating services.
  • the URL Recording and Rating module acts as an external interface to the URL rating database module. No modules or databases other than the main control module are directly connected. Access the database through the URL record and rating module.
  • the database is only connected to the UKL record and rating module and the main control module.
  • the URL record and rating module is connected to the following modules: the main control module; the URL rating database module; the decision module, the decision result is introduced to rank the associated UL score; the learning module, the learning process may refer to the database data; Hazardous Content Identification Module: During the identification process, it may be necessary to use URL data in the database. An example is: If a subtitle is superimposed on the video, telling the viewer to visit a URL, such as an illegal website, is also to be identified and controlled.
  • the filter node 3 ⁇ 4 includes:
  • the harmful content recording module is respectively connected to the I frame detection/decoding module and the decision module, and the decision module triggers disconnection of the video code stream, and starts the harmful content recorded by the harmful content recording module to record the harmful content;
  • the window length TW can be specified by a human monitor;
  • a video content recording module configured to record a video stream of a monitor for a specified period of time and stored in the recorded content storage module; the length of the recorded time window TW can be specified by a human monitor;
  • the harmful content recording module and the video content recording module are combined and set as one recording module
  • the recording content storage module is connected to the harmful content recording module and the video content recording module (ie, the recording module) for storing the recorded harmful content.
  • the filter node also includes:
  • the harmful content learning module is configured to connect the recorded content storage module, and when the automatic recognition sub-module and the monitor's recognition result of the content are inconsistent and finally execute the harmful judgment result of the monitor, learn the harmful content and learn The results are added to the harmful content database.
  • the harmful content learning module correspondingly includes:
  • An image learning unit connected to the harmful image database, for learning harmful images and adding the learning results to the harmful image database;
  • Superimposed text/symbol learning unit connected to the harmful superimposed text/symbol database, used to learn harmful superimposed text/symbols and add learning results to the harmful superimposed text/symbol database;
  • the face learning unit connected to the face database, is used to learn the face image and add the learning result to the face database.
  • the filtering node also includes the following structure:
  • the operation interface module is used for inputting relevant parameters or operation instructions; providing an operation interface for the human monitor, including a user graphical interface and a command line.
  • a feature network module is coupled between the interface module and the harmful image database for inputting/adjusting a feature network model and/or an event feature template to the harmful image recognition unit.
  • the parameter setting module is connected between the operation interface module and the scene segmentation module, and is used for inputting/adjusting relevant parameters required for scene segmentation in the scene segmentation module.
  • a decision rule setting module connected between the operation interface module and the decision module, for inputting/adjusting a decision rule of the control signal to the decision module;
  • a rating rule setting module connected between the operation interface module and the URL rating database, for inputting/adjusting a rating rule into the URL rating database
  • the main control module is respectively connected to any other module, sub-module or unit in the filtering node, and the module is a central module of the filtering node, and functions to control all other modules, sub-modules or units;
  • the log reporting module is respectively connected to any other module, sub-module or unit in the filtering node, and is used for logging and generating the result of the running state of the node, the event that occurs, and the result of the content filtering.
  • An external control module is connected to the main control module for performing data/signal interaction with an external control device. Because the node is deployed on the same network location as other media devices such as the media gateway in the network location, even in the physical device configuration, it can be implemented in the same physical device as the media gateway. Therefore, it is likely to accept the control of an external control device such as a gateway controller, and report information to an external device.
  • the communication protocol used for control commands and data reporting may be H.248MGCP.
  • control instruction module is connected between the operation interface module and the main control module for accepting instructions of the human monitor, such as cutting off harmful video code streams, replacing, initiating or prohibiting the filtering function based on the harmless code stream, restarting
  • the node and the like may be disposed in the control instruction module;
  • the filtering node of the present invention can be deployed on the network, and the network location is not strictly specified. In fact, it can be deployed at any network location between the content source and the user terminal, as long as the media stream to be filtered passes through the network location. In extreme cases, it can be deployed on the user terminal, which is equivalent to a content filtering subsystem built into the terminal.
  • video code stream is encrypted and does not affect the implementation of the technical solution of the present invention.
  • the encryption of the video stream has the following two possibilities:
  • the specific grading standard of the harmful content and the corresponding identification standard are determined according to the actual application scenario, and the specific standard or the identification method does not limit the protection scope of the present invention.
  • the obtained intra-coded frame to be detected may be completely decoded, and the framed image of the intra-coded frame may be identified as containing harmful content, and if so, the playback of the video stream is cut off; Play the video stream.
  • the technical solution provided by the embodiment of the present invention can also be used for recognizing other video content, for example, a sports game highlight shot, such as a soccer shot lens, a basketball long-range hit, a dunk, and the like, for the purpose of identifying the video. Fragments are stored and recorded; images related to specific people are identified from news programs for archiving; videos automatically recorded by electronic eyes (ie cameras installed at major intersections) used in the transportation system are identified, and violations are identified and identified The number of the illegal vehicle; identifies a specific story in the TV program, such as a Harry Potter movie, once it is recognized that the IPTV user can be notified to watch. It is apparent that those skilled in the art can make various modifications and variations to the invention without departing from the spirit and scope of the invention. Thus, it is intended that the present invention cover the modifications and the modifications of the invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Databases & Information Systems (AREA)
  • Computer Security & Cryptography (AREA)
  • Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

La présente invention concerne un procédé et un noeud de filtrage de flux vidéo dans la communication multimédia. Le procédé comprend l'obtention de l'intra-trame à partir du flux vidéo et le décodage de l'intra-trame; l'identification de contenu nuisible dans l'intra-trame décodée pour déterminer si la lecture du flux vidéo doit être stoppée. Le noeud comporte un module de retardement de flux vidéo, un module de commutation, et un module de détection/décodage de trame I, un module d'identification de contenu et un module de détermination.
PCT/CN2007/001463 2006-04-30 2007-04-29 Procédé et noeud de filtrage de flux vidéo WO2007128234A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CNB2006100790231A CN100490532C (zh) 2006-04-30 2006-04-30 一种视频码流过滤方法和过滤节点
CN200610079023.1 2006-04-30

Publications (1)

Publication Number Publication Date
WO2007128234A1 true WO2007128234A1 (fr) 2007-11-15

Family

ID=38076911

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2007/001463 WO2007128234A1 (fr) 2006-04-30 2007-04-29 Procédé et noeud de filtrage de flux vidéo

Country Status (2)

Country Link
CN (2) CN100490532C (fr)
WO (1) WO2007128234A1 (fr)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106550247A (zh) * 2016-10-31 2017-03-29 吴俊� 广播电视的监控方法
CN110020252A (zh) * 2017-12-30 2019-07-16 惠州学院 基于片尾内容的识别有害视频的方法及其系统
CN114143614A (zh) * 2021-10-25 2022-03-04 深蓝感知(杭州)物联科技有限公司 一种基于视频帧时延检测的网络自适应传输方法与装置
US11412303B2 (en) 2018-08-28 2022-08-09 International Business Machines Corporation Filtering images of live stream content
CN116109990A (zh) * 2023-04-14 2023-05-12 南京锦云智开软件有限公司 一种视频的敏感违规内容检测系统

Families Citing this family (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101557510A (zh) * 2008-04-09 2009-10-14 华为技术有限公司 视频编码处理方法、系统及装置
CN101982981B (zh) * 2010-11-12 2012-02-01 福州大学 数字电视传输流的分类检测装置
CN102073676A (zh) * 2010-11-30 2011-05-25 中国科学院计算技术研究所 一种网络色情视频实时检测方法和系统
CN102801956B (zh) * 2012-04-28 2014-12-17 武汉兴图新科电子股份有限公司 一种网络视频监控装置及方法
CN103106251B (zh) * 2013-01-14 2016-08-03 冠捷显示科技（厦门）有限公司 一种过滤显示设备不能播放的媒体文件的系统及过滤方法
CN104254002B (zh) * 2013-06-25 2018-01-12 上海尚恩华科网络科技股份有限公司 一种用于多地多频道的实时广告监播系统与方法
CN103596016B (zh) * 2013-11-20 2018-04-13 韩巍 一种多媒体视频数据处理方法和装置
US9742827B2 (en) * 2014-01-02 2017-08-22 Alcatel Lucent Rendering rated media content on client devices using packet-level ratings
CN104834689B (zh) * 2015-04-22 2019-02-01 上海微小卫星工程中心 一种码流类型快速识别方法
CN106708949A (zh) * 2016-11-25 2017-05-24 成都三零凯天通信实业有限公司 一种视频有害内容识别方法
KR102384878B1 (ko) * 2016-12-19 2022-04-11 삼성전자주식회사 동영상을 필터링하는 방법 및 장치
US10349126B2 (en) 2016-12-19 2019-07-09 Samsung Electronics Co., Ltd. Method and apparatus for filtering video
CN110019946A (zh) * 2017-12-30 2019-07-16 惠州学院 一种识别有害视频的方法及其系统
CN110109952A (zh) * 2017-12-30 2019-08-09 惠州学院 一种识别有害图片的方法及其系统
CN110020258A (zh) * 2017-12-30 2019-07-16 惠州学院 一种基于近似图的url路径识别有害图片的方法及系统
CN109993036A (zh) * 2017-12-30 2019-07-09 惠州学院 一种基于用户id识别有害视频的方法及其系统
CN109089126B (zh) * 2018-07-09 2021-04-27 武汉斗鱼网络科技有限公司 一种视频分析方法、装置、设备及介质
CN115734026A (zh) * 2021-08-31 2023-03-03 北京字跳网络技术有限公司 视频的处理方法、装置及系统
CN113891120A (zh) * 2021-09-29 2022-01-04 广东省高峰科技有限公司 一种iptv服务端接入方法及其系统
CN118317128A (zh) * 2024-04-16 2024-07-09 联通视频科技有限公司 一种机顶盒终端安全监测系统

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1484450A (zh) * 2002-08-13 2004-03-24 快速数字信道变换
CN1551624A (zh) * 2003-05-15 2004-12-01 �ձ������ȷ湫˾ 广播节目内容菜单建立装置和方法
CN1685724A (zh) * 2002-03-05 2005-10-19 汤姆森许可公司 在父母控制系统中选择性地访问节目的方法和设备

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1685724A (zh) * 2002-03-05 2005-10-19 汤姆森许可公司 在父母控制系统中选择性地访问节目的方法和设备
CN1484450A (zh) * 2002-08-13 2004-03-24 快速数字信道变换
CN1551624A (zh) * 2003-05-15 2004-12-01 �ձ������ȷ湫˾ 广播节目内容菜单建立装置和方法

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106550247A (zh) * 2016-10-31 2017-03-29 吴俊� 广播电视的监控方法
CN106550247B (zh) * 2016-10-31 2019-06-07 杭州天时亿科技有限公司 广播电视的监控方法
CN110020252A (zh) * 2017-12-30 2019-07-16 惠州学院 基于片尾内容的识别有害视频的方法及其系统
CN110020252B (zh) * 2017-12-30 2022-04-22 惠州学院 基于片尾内容的识别有害视频的方法及其系统
US11412303B2 (en) 2018-08-28 2022-08-09 International Business Machines Corporation Filtering images of live stream content
CN114143614A (zh) * 2021-10-25 2022-03-04 深蓝感知(杭州)物联科技有限公司 一种基于视频帧时延检测的网络自适应传输方法与装置
CN114143614B (zh) * 2021-10-25 2023-11-24 深蓝感知(杭州)物联科技有限公司 一种基于视频帧时延检测的网络自适应传输方法与装置
CN116109990A (zh) * 2023-04-14 2023-05-12 南京锦云智开软件有限公司 一种视频的敏感违规内容检测系统

Also Published As

Publication number Publication date
CN1968408A (zh) 2007-05-23
CN101317455A (zh) 2008-12-03
CN100490532C (zh) 2009-05-20

Similar Documents

Publication Publication Date Title
WO2007128234A1 (fr) Procédé et noeud de filtrage de flux vidéo
CN109348171B (zh) 一种视联网监控视频获取方法及装置
JP4601868B2 (ja) ネットワーク化した監視／管制システム及び方法
RU2321967C2 (ru) Способ маскирования ошибок в видеопоследовательностях
TWI458340B (zh) 用於多工視訊分量之發信號資料
CN110324580B (zh) 一种基于视联网的监控视频播放方法及装置
CN110545405B (zh) 一种基于视联网的视频传输方法及系统
CN109756789B (zh) 一种音视频数据包的丢包处理方法和系统
CN108965986B (zh) 一种视频录播方法和系统
JP6768951B2 (ja) ネットワーク分離装置およびこれを備える映像監視システム
CN110769310B (zh) 一种基于视联网的视频处理方法和装置
CN110012322B (zh) 一种视联网业务发起的方法和系统
CN114640655B (zh) 一种基于hls视频播放的安全视频调阅系统和方法
US20230045876A1 (en) Video Playing Method, Apparatus, and System, and Computer Storage Medium
CN111294639A (zh) 一种视频实时在线共享浏览防篡改的系统和方法
KR101857716B1 (ko) 망 분리 장치 및 이를 구비하는 영상 감시 시스템
CN108965930B (zh) 一种视频数据处理的方法和装置
CN108989833B (zh) 一种视频封面图像的生成方法及装置
CN109905627B (zh) 一种音视频流数据录制的方法和装置
CN110719247B (zh) 终端入网方法和装置
WO2007131445A1 (fr) Procédé, système et appareil pour censurer un flux de code vidéo
Go et al. Secure video transmission framework for battery-powered video devices
CN108632635B (zh) 一种基于视联网的数据处理方法和装置
CN110324667B (zh) 一种新型视频流的播放方法和系统
CN111447407A (zh) 一种监控资源的传输方法和装置

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 200780000398.7

Country of ref document: CN

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 07721036

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 07721036

Country of ref document: EP

Kind code of ref document: A1