CN115830508B - 5G message content detection method - Google Patents
5G message content detection method Download PDFInfo
- Publication number
- CN115830508B CN115830508B CN202211600092.8A CN202211600092A CN115830508B CN 115830508 B CN115830508 B CN 115830508B CN 202211600092 A CN202211600092 A CN 202211600092A CN 115830508 B CN115830508 B CN 115830508B
- Authority
- CN
- China
- Prior art keywords
- frame
- video
- violation
- frames
- illegal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title abstract description 12
- 230000011218 segmentation Effects 0.000 claims abstract description 59
- 238000012216 screening Methods 0.000 claims abstract description 41
- 238000000034 method Methods 0.000 claims description 26
- 238000004364 calculation method Methods 0.000 claims description 25
- 230000002159 abnormal effect Effects 0.000 claims description 16
- 230000003044 adaptive effect Effects 0.000 claims description 13
- 238000005192 partition Methods 0.000 claims description 12
- 238000001914 filtration Methods 0.000 claims 1
- 238000004891 communication Methods 0.000 abstract description 7
- 238000012544 monitoring process Methods 0.000 abstract description 4
- XOFYZVNMUHMLCC-ZPOLXVRWSA-N prednisone Chemical compound O=C1C=C[C@]2(C)[C@H]3C(=O)C[C@](C)([C@@](CC4)(O)C(=O)CO)[C@@H]4[C@@H]3CCC2=C1 XOFYZVNMUHMLCC-ZPOLXVRWSA-N 0.000 description 4
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 102000003712 Complement factor B Human genes 0.000 description 1
- 108090000056 Complement factor B Proteins 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of communication network monitoring, in particular to a 5G message content detection method. Firstly, acquiring video information in a 5G message; screening out characteristic frames in the video image; matching each frame of video image in the preliminary screening illegal video with the characteristic frame to obtain multi-frame illegal frames; when the characteristic frame is matched with different violation frames, carrying out self-adaptive segmentation to obtain self-adaptive segmentation size, segmenting the characteristic frame and the violation frames to obtain segmentation areas, and calculating area common factors corresponding to the segmentation areas in the characteristic frame; obtaining an overall common factor according to the regional common factors of each divided region; calculating the violation stability according to the fluctuation degree of the integral public factor; and identifying the violations of the video information according to the violation stability. The invention uses the characteristic frame and the illegal frame in the illegal video database to carry out the content identification and matching of the self-adaptive area size, and can also carry out illegal identification after the illegal video is clipped.
Description
Technical Field
The invention relates to the technical field of communication network monitoring, in particular to a 5G message content detection method.
Background
With the development and upgrade of communication technology, the former short message communication service is also upgraded into 5G message and is fully commercially available. The 5G message is more diversified than the prior short message communication, and supports the communication of various types of contents. The message brings convenience to people and also brings certain challenges, such as monitoring and management of bad messages, unlike the traditional text messages, only needs to monitor a small amount of text messages, the detection of various types of contents of the 5G message is more complex than the detection of the traditional text messages, especially the identification of the contents of videos in the 5G message is more difficult, and the provision of high-efficiency, low-delay, comprehensive and accurate bad message identification is the key for guaranteeing the user experience for improving the user experience.
When the video message in the 5G message is identified, the current common detection method for the video content is to match through the hash value of the video in order to meet the detection low delay, namely, the video content in the 5G message is generated into a corresponding hash value by utilizing a hash algorithm, then the hash value is compared with the hash value of the illegal video in the stored illegal video library to identify the illegal video, but the hash value is extremely sensitive to the content of the video, namely, the same illegal video is subjected to certain content editing to cause the hash value of the video to be different, so that the success rate of the illegal video identification is reduced, and the illegal video after editing is difficult to identify.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a 5G message content detection method, which adopts the following technical scheme:
acquiring video information in the 5G message; preliminary matching is carried out on the video information and the violation videos in the violation video database, so that a preliminary screening violation video is obtained;
screening out characteristic frames in the video images according to the information entropy of each frame of video image in the video information; matching each frame of video image in the preliminary screening illegal video with the characteristic frame to obtain multi-frame illegal frames; performing self-adaptive segmentation when the characteristic frame is matched with different illegal frames to obtain self-adaptive segmentation size;
dividing the feature frame and the violation frame based on the self-adaptive division size to obtain at least two division areas, and obtaining an area public factor corresponding to each division area in the feature frame according to the similarity of each division area in the feature frame and each division area in the violation frame; for the characteristic frames and any corresponding violation frames, obtaining overall common factors according to the regional common factors of each divided region;
calculating the violation stability according to the fluctuation degree of the integral public factor; and identifying the violation condition of the video information in the 5G message according to the violation stability.
Preferably, the calculation formula of the adaptive segmentation size is as follows:
wherein X is m,j The adaptive segmentation size is matched with the feature frame of the m frame and the corresponding violation frame of the j frame; exp is an exponential function with a natural constant as a base; d (D) j The Hamming distance of the hash value of the m-th frame characteristic frame and the j-th frame illegal frame is used; AVG (D) m,J ) The average value of Hamming distances of hash values of the m-th frame characteristic frame and all corresponding offending frames is obtained; NG (NG) m The number of pixel points in the feature frame of the m-th frame; NG'. j The number of pixel points in the j-th frame violation frame is the number of pixel points in the j-th frame violation frame;to round the symbol up.
Preferably, the obtaining the region common factor corresponding to each divided region in the feature frame according to the similarity between each divided region in the feature frame and each divided region in the violation frame includes:
selecting any segmentation area in the feature frame as a target segmentation area;
calculating the similarity between the target segmentation area and each segmentation area in the violation frame; selecting the maximum similarity as a region common factor corresponding to the target segmentation region;
the calculation formula of the similarity is as follows:
wherein XS mk,j′k The similarity between the kth divided area in the characteristic frame of the mth frame and the kth divided area in the violation frame of the jth frame is obtained; x is X m,j The adaptive segmentation size is matched with the feature frame of the m frame and the corresponding violation frame of the j frame; gmk i The gray value of the ith pixel point in the kth partition area in the mth frame characteristic frame is obtained; gj' k i The gray value of the ith pixel point in the kth partition area in the jth frame violation frame is used as the gray value of the ith pixel point; gmk i′ The gray value of the ith pixel point in the kth partition area in the mth frame characteristic frame corresponding to the ith pixel point in the eighth adjacent area is obtained；Gj′k i′ The gray value of the ith pixel point in the eighth adjacent area corresponding to the ith pixel point in the kth partition area in the jth frame of violation frame; exp is an exponential function that bases on a natural constant.
Preferably, the preliminary matching of the video information and the violation videos in the violation video database to obtain a preliminary screening violation video includes:
acquiring a hash value of the video information as a first hash value; acquiring a hash value of any illegal video in the illegal video database as a second hash value; and calculating the Hamming distance between the first Hash value and the second Hash value, and selecting the illegal video with the Hamming distance smaller than a preset first threshold value as the primary screening illegal video.
Preferably, the screening the feature frames in the video image according to the information entropy of each frame of the video image in the video information includes:
selecting any video image as a target image, acquiring information entropy corresponding to the target image, and calculating abnormal values of the target image relative to the front and rear frames of video images according to the information entropy corresponding to the target image;
and when the abnormal value is greater than or equal to a preset second threshold value, taking the target image as a characteristic frame in the video image.
Preferably, the calculating the abnormal value of the target image relative to the two frames of video images according to the information entropy corresponding to the target image includes:
acquiring information entropy of a front frame video image and a rear frame video image corresponding to a target image; calculating the average value of the information entropy of the target image and the front and rear two frames of video images, and taking the average value as the average value of the information entropy; calculating the absolute value of the difference value between the information entropy of the target image and the information entropy mean value; the ratio of the absolute value of the difference value and the information entropy of the target image is an abnormal value corresponding to the target image.
Preferably, the matching the video image of each frame in the preliminary screening violation video with the feature frame to obtain a multi-frame violation frame includes:
acquiring a hash value of any characteristic frame as a third hash value; selecting any prescreened illegal video as a target video, and acquiring a hash value of any video image in the target video as a fourth hash value; calculating the Hamming distance of the third Hamming value and the fourth Hamming value, and taking a video image in the target video corresponding to the minimum Hamming distance as a violation frame;
and for any characteristic frame, obtaining a plurality of violation frames corresponding to the preliminary screening violation videos to obtain a plurality of frame violation frames.
Preferably, the calculating the violation stability according to the fluctuation degree of the overall common factor includes:
and for any feature frame, acquiring the integral common factors of all violation frames corresponding to the feature frame, and calculating the difference value of the integral common factors corresponding to adjacent violation frames, wherein the sum of the difference values is used as the violation stability.
Preferably, the identifying the violation condition of the video information in the 5G message according to the violation stability includes:
when the violation stability corresponding to any feature frame is greater than or equal to a preset third threshold, the corresponding video information is a violation video; otherwise, the corresponding video information is normal video.
Preferably, the overall common factor is: the average value of the region common factors of the respective divided regions.
The embodiment of the invention has at least the following beneficial effects:
firstly, acquiring video information in a 5G message; preliminary matching is carried out on the video information and the illegal video in the illegal video database, so that a preliminary screening illegal video is obtained; although the whole video of the illegal video is changed from the original illegal video before being clipped after being clipped to a certain extent, the illegal video still has certain relativity, so that the illegal video in the illegal video database is preliminarily matched. Feature frames in the video images are screened out, and the extracted feature frames are used for matching with illegal videos in the illegal video database, so that the calculated amount of video matching can be reduced, and the matching accuracy can be improved. Matching each frame of video image in the preliminary screening illegal video with the characteristic frame to obtain multi-frame illegal frames; based on the Hamming distance between frames, when the feature frames are matched with different violation frames, self-adaptive segmentation is carried out to obtain self-adaptive segmentation sizes, the feature frames and the violation frames are segmented to obtain segmentation areas, and because each feature frame and the corresponding violation frame have the same partial content, but the same content is possibly not in the same position, the self-adaptive segmentation sizes based on the Hamming distance are carried out on each feature frame and each violation frame in the corresponding violation frame set, and the self-adaptive segmentation sizes suitable for the current matching are segmented to facilitate the follow-up violation identification of the feature frames. Obtaining region common factors corresponding to each divided region in the feature frame according to the similarity of each divided region in the feature frame and each divided region in the violation frame; for the characteristic frames and any corresponding violation frames, obtaining overall common factors according to the regional common factors of each divided region; calculating the violation stability according to the fluctuation degree of the integral public factor; and identifying the violation condition of the video information in the 5G message according to the violation stability. The invention extracts the characteristic frame of the video information in the 5G message, and performs self-adaptive region content identification and matching with the offending frame in the offending video database, so that compared with the prior art of the content identification of the video message in the 5G message, the offending identification can be performed after the offending video is clipped, and the offending video identification has larger strength.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for detecting content of a 5G message according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purposes, the following detailed description refers to specific embodiments, structures, features and effects of a 5G message content detection method according to the present invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The embodiment of the invention provides a specific implementation method of a 5G message content detection method, which is suitable for a violation identification scene of video information in a 5G message. In order to solve the problem that the hash value is extremely sensitive to the content of the video, the same illegal video can cause different hash values of the video after editing certain content, so that the success rate of identifying the illegal video is reduced, and the problem that the illegal video after editing is difficult to identify is solved. According to the method, the characteristic frames of the video information in the 5G message are extracted, the characteristic frames are subjected to self-adaptive region-size content identification and matching with the illegal frames in the illegal video database, and compared with the content identification of the video message in the existing 5G message, the illegal identification can be performed after the illegal video is clipped, so that the illegal video identification has higher strength.
The following specifically describes a specific scheme of a 5G message content detection method provided by the present invention with reference to the accompanying drawings.
Referring to fig. 1, a flowchart illustrating steps of a method for detecting 5G message content according to an embodiment of the present invention is shown, the method includes the following steps:
step S100, obtaining video information in a 5G message; and performing preliminary matching on the video information and the violation videos in the violation video database to obtain a preliminary screening violation video.
The method aims to achieve the purpose of reducing workload when identifying the violation condition of the video information in the 5G message, and firstly, performs preliminary matching on the video information in the 5G message and the violation video in the violation video database to obtain a preliminary screening violation video corresponding to the current video information. In the embodiment of the invention, the illegal video database is a illegal video database containing unrepeated illegal contents through manual identification and screening.
The preliminary matching method comprises the following steps: acquiring a hash value of video information as a first hash value; acquiring a hash value of any illegal video in the illegal video database as a second hash value; and calculating the Hamming distance between the first Hash value and the second Hash value, and selecting the illegal video with the Hamming distance smaller than a preset first threshold value as the primary screening illegal video. In the embodiment of the present invention, the value of the first threshold is preset to be 10, and in other embodiments, the practitioner may adjust the value according to the actual situation. The method comprises the steps of generating a hash value of video information in 5G information of content to be identified, calculating Hamming distances with all offending videos in an offending video database by using the hash value, and selecting offending videos with smaller Hamming distances as preliminary screening offending videos of the 5G information to be identified. The primary matching is performed to change the hash value of the whole video corresponding to the illegal video after the illegal video is clipped to a certain extent, but the hash value still has certain relativity, so that the Hamming distance of the hash value is utilized to perform primary screening of the 5G message to be identified, and the primary screening illegal video corresponding to the video information in the 5G message is obtained.
Step S200, screening out characteristic frames in the video images according to the information entropy of each frame of video image in the video information; matching each frame of video image in the preliminary screening illegal video with the characteristic frame to obtain multi-frame illegal frames; and performing self-adaptive segmentation when the characteristic frame is matched with different illegal frames to obtain the self-adaptive segmentation size.
The process of carrying out illegal content identification on the video message in the 5G message comprises the steps of carrying out feature frame extraction by utilizing the video message in the 5G message, then carrying out illegal frame acquisition by utilizing the hash value and hamming distance of each extracted feature frame and the video image of the illegal video in the illegal video database, carrying out self-adaptive size region division on each feature frame and each frame of the illegal video in the corresponding illegal frame, further, obtaining a region common factor by carrying out similarity calculation on each divided region of the feature frame and each divided region of the illegal frame in the corresponding illegal frame set, then obtaining the integral common factor of each feature frame and the corresponding illegal frame by utilizing the region common factor, and finally identifying whether the current feature frame is illegal or not by the stability of the integral common factor, thereby identifying whether the video in the 5G message is the illegal video or not.
Firstly, screening out characteristic frames in video images according to the information entropy of each frame of video image in video information, namely extracting the characteristic frames by utilizing video information in 5G information.
The feature frames of the video information in the 5G message refer to video images with larger variability from other frame video images in all frame video images of the video message of the 5G message, the video images are not always in strong relation with the whole video, and the feature frames are always ignored when general clipping software clips illegal videos in a database, so that the feature frames are extracted to match the illegal videos in the illegal video database, the calculated amount of video matching can be reduced, and the matching accuracy can be improved. Taking video information in a certain 5G message as an example, the extraction mode of a characteristic frame is to select any video image as a target image, obtain information entropy corresponding to the target image, and calculate an abnormal value of the target image relative to the video images of the front frame and the rear frame according to the information entropy corresponding to the target image; and when the abnormal value is greater than or equal to a preset second threshold value, taking the target image as a characteristic frame in the video image. In the embodiment of the present invention, the value of the second threshold is preset to be 0.45, and in other embodiments, the practitioner can adjust the value according to the actual situation. The feature frame is obtained specifically:
firstly, calculating the whole information quantity of each frame of video image of the video information, namely calculating the information entropy of each frame of video image. Taking an nth frame of video image in video information as an example, the corresponding information entropy E n The calculation mode of (a) is as follows:
wherein p is g The occurrence frequency of the pixel point with the gray value g in the nth frame of video image is obtained; log is a logarithmic function. The frequency of occurrence is calculated by dividing the number of times of occurrence of the pixel point with the gray value g in the video image of the nth frame by the number of all the pixel points of the nth frame, wherein g is 0,255]. It should be noted that the calculation formula of the information entropy is a well-known technology of those skilled in the art, and will not be described herein.
Further, performing difference value calculation based on continuous frame information quantity on continuous frames to screen characteristic frames in a plurality of video images, namely calculating abnormal values of a target image relative to the video images of the front frame and the rear frame according to information entropy corresponding to the target image; and when the abnormal value is greater than or equal to a preset second threshold value, taking the target image as a characteristic frame in the video image.
The calculation formula of the outlier is as follows:
wherein ΔE is n An outlier for the n-th frame video image; e (E) n Information entropy of the nth frame video image;is the average value of the information entropy of the nth frame of video image and the front and back frames of video images.
The information entropy reflects the whole information quantity of the video image; n is E [2, N]Where N is the maximum number of frames corresponding to the current video information. The logic of the calculation formula of the outlier is as follows: firstly, taking an nth frame of video image as a basis, averaging the whole information quantity of the nth frame of video image and the whole information quantity of two continuous frames of video images before and after, and then obtaining a difference with the whole information quantity of the nth frame of video image, wherein the larger the difference is, the more abnormal the video image of the nth frame is reflected, and the larger the abnormal value of the corresponding video image is; conversely, the smaller the difference, the more normal the frame of video image is reflected, the corresponding video imageThe smaller the outlier of (c). In the video information, generally, the difference between the video of each frame and the video of the successive frame is not large, so that the average value of the total information amount corresponding to the video image of the successive frame before and after the video image of the nth frame is correspondingly small from the average value of the total information amount of the video image of the nth frame. However, when a special frame is encountered, the difference between the frame and the video images of the previous and subsequent frames is strong, so that the average value of the total information amount of the corresponding continuous frame video and the average value of the total information amount of the nth frame video are greatly different, the frame video is quantized, and then the ratio of the frame video to the total information amount of the nth frame video is calculated to represent the abnormal value delta E of the nth frame video n Abnormal value delta E n The larger the relation between the nth frame video image and the successive frame video images is reflected, the smaller the relation is, and conversely, the outlier delta E is n The smaller the relation between the nth frame video image and the preceding and following successive frame video images is reflected to be larger. When the anomaly value delta E n When larger, the nth frame video image can be considered as a feature frame. That is, when the abnormal value is greater than or equal to a preset second threshold, the nth frame of video image is considered as a characteristic frame.
Screening all frame video images of the video information in the 5G message to obtain all M characteristic frames M E [1, N ] in the video images in all the 5G message; where N is the total number of video images corresponding to the video information.
After screening out the characteristic frames in the video images, matching each frame of video image in the preliminary screening illegal video with the characteristic frames to obtain multi-frame illegal frames.
Firstly, hash value acquisition is carried out on all frames in all preliminary screening illegal videos in the preliminarily screened illegal video database and M characteristic frames in the extracted 5G message. It should be noted that, the method for obtaining the hash value is a well-known technique of those skilled in the art, and will not be described herein.
And then, carrying out Hash value calculation on Hash distances between any one of M feature frames and all frame video images of the preliminary screening violation videos in the preliminary screening violation video library, and selecting one violation frame with the smallest Hash distance with each feature frame in each preliminary screening violation video as a violation frame set of each feature frame, namely, each feature frame has a violation frame set formed by the violation frames with the smallest Hash distances with the violation frames from different preliminary screening violation videos.
Obtaining a plurality of violation frames corresponding to each characteristic frame, namely: obtaining a hash value of any characteristic frame as a third hash value; selecting any prescreened illegal video as a target video, and acquiring a hash value of any video image in the target video as a fourth hash value; calculating the Hamming distance of the third Hamming value and the fourth Hamming value, and taking a video image in the target video corresponding to the minimum Hamming distance as a violation frame; and for any characteristic frame, obtaining a plurality of violation frames corresponding to the preliminary screening violation videos to obtain a plurality of frame violation frames.
After obtaining the violation frame corresponding to each feature frame, the segmentation size of the feature frame and the matched violation frame is self-adapted according to the hamming distance between frames, namely, the self-adaptive segmentation is performed when the feature frame is matched with different violation frames based on the hamming distance between frames, so as to obtain the self-adaptive segmentation size.
A hamming distance is utilized to obtain a relatively similar violation frame set corresponding to each of M feature frames of the 5G message, each frame of the violation frame set corresponding to each feature frame is likely to have partial identical content, but the identical content is likely to be different from the identical position, so that the hamming distance-based adaptive segmentation size is carried out on each feature frame and each violation frame of the corresponding violation frame set, at least two segmentation areas are obtained by utilizing the adaptive segmentation size by segmenting each feature frame and each violation frame of the corresponding violation frame set, and then the content identification and matching are carried out on different areas in the feature frames and each violation frame.
Taking the mth feature frame in the feature frames as an example, a calculation formula of the adaptive segmentation size of the mth feature frame and the j-th violation frame in the corresponding violation frame set is as follows:
wherein X is m,j The adaptive segmentation size is matched with the feature frame of the m frame and the corresponding violation frame of the j frame; exp is an exponential function with a natural constant as a base; d (D) j The Hamming distance of the hash value of the m-th frame characteristic frame and the j-th frame illegal frame is used; AVG (D) m,J ) The average value of Hamming distances of hash values of the m-th frame characteristic frame and all corresponding offending frames is obtained; NG (NG) m The number of pixel points in the feature frame of the m-th frame; NG'. j The number of pixel points in the j-th frame violation frame is the number of pixel points in the j-th frame violation frame;to round the symbol up.
Wherein X is m,j Representing the side length of each divided area when the content identification and matching of the mth characteristic frame and the jth offending frame in the corresponding offending frame set should be performed with the regional division, namely the size of the divided area is X m,j ×X m,j ，j∈[1,J]Wherein J is the total number of all offending frames in the offending frame set corresponding to the mth feature frame.
The calculation formula logic of the self-adaptive segmentation size: hamming distance D of hash value of mth frame characteristic frame and jth frame offending frame j The greater the hamming distance relative to the entire offending frame set, the |d j -AVG(D m,J ) The larger the I is, the less the content of the feature frame of the m frame is compared with the content of the violation frame of the j frame, so that the region division of the feature frame of the m frame is smaller compared with the rest of violation frames, and the matching result in the subsequent matching is more accurate. Otherwise, the more the content of the characteristic frame of the mth frame is identical with the illegal frame of the jth frame, the |D j -AVG(D m,J ) The smaller the i is, the more easily it can be recognized that the same content is matched even if the area is large, so that the divided area thereof is enlarged to reduce the calculation amount in the subsequent matching. And then the negative exponential function taking the natural constant as the base is used for inverting the natural constant, so that smaller parameters are obtained when the Hamming distance is larger, larger parameters are obtained when the Hamming distance is smaller,then the average value of all pixel points of the mth characteristic frame and the jth offending frame is usedAs a basis, the adaptive segmentation size when content matching and identification are performed on the mth feature frame and the jth offending frame in the following is obtained by multiplying the above-described obtaining parameters.
And calculating all the characteristic frames to obtain the self-adaptive segmentation size corresponding to all the characteristic frames and each frame of the corresponding violation frame set. The Hamming distance between each characteristic frame and the hash value of each offending frame in the corresponding offending frame set is utilized, and the self-adaptive segmentation size in the subsequent content matching and recognition is obtained.
Step S300, dividing the feature frame and the violation frame based on the self-adaptive division size to obtain at least two division areas, and obtaining an area common factor corresponding to each division area in the feature frame according to the similarity of each division area in the feature frame and each division area in the violation frame; and for the characteristic frames and any corresponding violation frames, obtaining the integral common factors according to the regional common factors of the partitioned regions.
In order to accurately identify the illegal contents in the feature frames, each feature frame and the corresponding illegal frame are segmented by utilizing the obtained adaptive segmentation size to obtain a plurality of at least two segmentation areas. And carrying out matching calculation on the region public factors by using each divided region in the feature frame and each divided region in the corresponding violation frame, namely obtaining the region public factors corresponding to each divided region in the feature frame according to the similarity of each divided region in the feature frame and each divided region in the violation frame, and further carrying out calculation on the whole public factors of each feature frame by using the region public factors.
Dividing the feature frame and the violation frame based on the self-adaptive division size to obtain at least two division areas, and obtaining an area common factor corresponding to each division area in the feature frame according to the similarity of each division area in the feature frame and each division area in the violation frame, wherein the m-th feature frame and the m-th feature frame are used specificallyj offending frames are exemplified, whose region common factor B j The calculation mode of (a) is as follows:
firstly, the m-th characteristic frame and the j-th violation frame are subjected to region division according to the corresponding self-adaptive division size, and m 'division regions corresponding to the m-th characteristic frame after division and j' division regions corresponding to the j-th violation frame can be obtained. And calculating the similarity between each divided area in the mth characteristic frame and each divided area in the j offending frame. Namely, selecting any segmentation area in the feature frame as a target segmentation area, and calculating the similarity between the target segmentation area and each segmentation area in the violation frame.
The calculation formula of the similarity is as follows:
wherein XS mk,j′k The similarity between the kth divided area in the characteristic frame of the mth frame and the kth divided area in the violation frame of the jth frame is obtained; x is X m,j The adaptive segmentation size is matched with the feature frame of the m frame and the corresponding violation frame of the j frame; gmk i The gray value of the ith pixel point in the kth partition area in the mth frame characteristic frame is obtained; gj' k i The gray value of the ith pixel point in the kth partition area in the jth frame violation frame is used as the gray value of the ith pixel point; gmk i′ The gray value of the ith pixel point in the kth partition area in the mth frame feature frame corresponds to the ith pixel point in the eighth adjacent area; gj' k i′ The gray value of the ith pixel point in the eighth adjacent area corresponding to the ith pixel point in the kth partition area in the jth frame of violation frame; exp is an exponential function that bases on a natural constant.
Wherein the gray value Gm1 i ∈[0,255]，i∈[1,X m,j ×X m,j ]，Gm1 i′ ∈[0,255]，i′∈[1,8]，Gj′1 i′ ∈[0,255]，Gj′1 i ∈[0,255]。
The logic of the calculation formula of the similarity is as follows: the feature frame of the m frame shares X with the first divided area of the violation frame of the j frame m,j ×X m,j Each pixel point is utilizedAnd calculating the difference value of the gray value of the pixel point at the corresponding position by the pixel point at the position so as to represent the similarity of the two divided areas, wherein if the pixels of the two divided areas are more similar, the similarity value is smaller. However, there is a certain chance that although the mth feature frame is the same as the ith pixel in the first region of the jth offending frame, the contribution to the whole region cannot be provided, so that the pixel in each position is constrained by equally calculating the average value of the difference values of the pixels in the 8 adjacent regions around the pixel in each position, if the pixels in not only the ith pixel but also the pixels in the 8 adjacent regions around the two divided regions are similar, it is explained that the pixel in the ith pixel is not only similar, and the contribution provided by the pixel in the ith pixel when the similarity calculation is performed on the two regions is not fortuitous, in a specific constraint manner that the average value of the differences of the pixels in the 8 adjacent regions around the pixel in the ith pixel is utilizedThe difference between the pixel points at the ith position in the two divided areas is averaged, so that even if the pixel points at the ith position in the two divided areas show small accidental difference, the influence of the average value of the difference values of the pixel points in the 8 neighborhood around the pixel points is larger. And then for all X's in both partitions in this manner m,j ×X m,j And calculating the pixel points to obtain the similarity, wherein the larger the value of the similarity is, the larger the similarity of the two divided areas is, and otherwise, the smaller the value of the similarity is, the smaller the similarity of the two divided areas is.
By using the method, the similarity of the first divided area in the m-th characteristic frame and each divided area in the j-th violation frame is calculated, and the similarity of the first divided area in the m-th characteristic frame and each divided area in the j-th violation frame can be obtained. And selecting the maximum similarity, namely representing the highest similarity, which is the region common factor of the first divided region in the mth characteristic frameB 1 Namely, after calculating the similarity between the target segmentation area and each segmentation area in the violation frame, selecting the maximum similarity as an area common factor corresponding to the target segmentation area.
The region common factor of all the divided regions in the mth feature frame can be obtained by calculation.
Further, for the feature frame and any corresponding violation frame, an overall common factor is obtained according to the regional common factor of each divided region, and the method is specific: the average value of the regional common factors of the divided regions is the integral common factor. For example, for each divided area obtained when the feature frame of the mth frame is matched with the corresponding violation frame of the jth frame, calculating the area common factor of each divided area, and further calculating the integral common factor when the feature frame of the mth frame is matched with the violation frame of the jth frame.
The calculation formula of the integral common factor is as follows:
wherein Bm j The integral common factor when the feature frame of the m frame and the illegal frame of the j frame are matched; m' is the number of divided regions of the m-th frame feature frame; b (B) m″ And when the m-th frame characteristic frame is matched with the j-th frame illegal frame, the region common factor of the m-th region in the m-th frame characteristic frame.
The logic of the calculation formula of the integral common factor is as follows: when the feature frame of the m frame and the violation frame of the j frame are matched, the common factors of all the divided areas in the feature frame are averaged, the larger the average value is, the more similar each divided area in the feature frame of the m frame is reflected to the divided area in the violation frame of the j frame in the matching process, and the similarity degree is higher, otherwise, when the average value is smaller, the more similar each divided area in the feature frame of the m frame is not reflected to the divided area in the violation frame of the j frame in the matching process, and the similarity degree between the divided areas is relatively lower.
And obtaining the integral common factors of the mth feature frame and each frame of the violation frames in the violation frame set corresponding to the mth feature frame through calculation, namely, how many violation frames are in the violation frame set, and how many integral common factors are in the mth feature frame.
Step S400, calculating the violation stability according to the fluctuation degree of the integral public factor; and identifying the violation condition of the video information in the 5G message according to the violation stability.
After the overall common factor is obtained, detecting the illegal video information in the 5G message by utilizing the overall common factor. And obtaining J integral common factors of the m feature frames and J frame violation frames in the corresponding violation frame set, and then detecting whether the m feature frames are violation frames or not through the J integral common factors. Calculating violation stability according to the fluctuation degree of the integral common factors corresponding to the J violation frames, and specifically: and for any feature frame, acquiring the integral common factors of all violation frames corresponding to the feature frame, and calculating the difference value of the integral common factors corresponding to adjacent violation frames, wherein the sum of the difference values is used as the violation stability corresponding to the feature frame.
The calculation formula of the violation stability is as follows:
wherein delta m The violation stability corresponding to the feature frame of the mth frame is obtained; bm j The integral common factor when the feature frame of the m frame and the illegal frame of the j frame are matched; bm j-1 The integral common factor when the feature frame of the m frame is matched with the violation frame of the j-1 frame; j is the number of offending frames corresponding to the m-frame feature frames.
Because if the mth feature frame is the offending frame, the m feature frame must have quite large similarity with one of J offending frames in the corresponding offending frame set, namely the integral common factor corresponding to the m feature frame is quite large, and the similarity with the rest J-1 offending video frames is quite small, namely the common factor corresponding to the m feature frame is quite small, the integral common factors of adjacent offending frames in all J offending frames are differenced, and J integral common factors are carried out by carrying out difference value accumulationCalculation of stability of co-factor when stability delta is violated m The closer to 0, the less likely it is that the mth feature frame is a offending frame, when offending stability delta m The larger the probability that the mth feature frame is the offending frame is explained to be greater.
And calculating the violation stability of the integral common factor for each of M feature frames corresponding to the video information in the 5G message, so that the corresponding violation stability of all M feature frames can be obtained.
And using the corresponding violation stabilities of all M feature frames to determine whether the video information in the 5G message is a violation video, namely. After obtaining the violation stability, identifying the violation condition of the video information in the 5G message according to the violation stability, and specifically: when the violation stability corresponding to any feature frame is greater than or equal to a preset third threshold, the corresponding video information is a violation video; otherwise, the corresponding video information is normal video. In the embodiment of the present invention, the value of the third threshold is preset to be 0.4, and in other embodiments, the practitioner can adjust the value according to the actual situation.
And (3) whether the video information in the 5G message is illegal or not is completely identified according to the corresponding illegal stability of the M feature frames. Further, processing the offensive video in the identified 5G message, for example, judging whether the video information content in the 5G message belongs to the offensive video by utilizing the offensive stability of the integral common factor of the feature frame, then shielding the 5G message containing the offensive video, identifying a sender according to the sending mobile phone number of the 5G message, and tracking the sender.
In summary, the present invention relates to the technical field of communication network monitoring. Firstly, acquiring video information in a 5G message; preliminary matching is carried out on the video information and the illegal video in the illegal video database, so that a preliminary screening illegal video is obtained; screening out characteristic frames in the video image; matching each frame of video image in the preliminary screening illegal video with the characteristic frame to obtain multi-frame illegal frames; based on the Hamming distance between frames, carrying out self-adaptive segmentation when the characteristic frames are matched with different violation frames to obtain self-adaptive segmentation sizes, segmenting the characteristic frames and the violation frames to obtain segmentation areas, and obtaining area common factors corresponding to all segmentation areas in the characteristic frames according to the similarity of each segmentation area in the characteristic frames and each segmentation area in the violation frames; for the characteristic frames and any corresponding violation frames, obtaining overall common factors according to the regional common factors of each divided region; calculating the violation stability according to the fluctuation degree of the integral public factor; and identifying the violation condition of the video information in the 5G message according to the violation stability. The invention extracts the characteristic frame of the video information in the 5G message, and performs self-adaptive region content identification and matching with the offending frame in the offending video database, compared with the prior 5G message, the method and the device have more accurate content identification and greater strength on offending video identification.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (7)
1. A method for detecting content of a 5G message, the method comprising the steps of:
acquiring video information in the 5G message; preliminary matching is carried out on the video information and the violation videos in the violation video database, so that a preliminary screening violation video is obtained;
screening out characteristic frames in the video images according to the information entropy of each frame of video image in the video information; matching each frame of video image in the preliminary screening illegal video with the characteristic frame to obtain multi-frame illegal frames; performing self-adaptive segmentation when the characteristic frame is matched with different illegal frames to obtain self-adaptive segmentation size;
dividing the feature frame and the violation frame based on the self-adaptive division size to obtain at least two division areas, and obtaining an area public factor corresponding to each division area in the feature frame according to the similarity of each division area in the feature frame and each division area in the violation frame; for the characteristic frames and any corresponding violation frames, obtaining overall common factors according to the regional common factors of each divided region;
calculating the violation stability according to the fluctuation degree of the integral public factor; identifying the violation condition of the video information in the 5G message according to the violation stability;
obtaining the region common factor corresponding to each divided region in the feature frame according to the similarity of each divided region in the feature frame and each divided region in the violation frame, wherein the method comprises the following steps:
selecting any segmentation area in the feature frame as a target segmentation area;
calculating the similarity between the target segmentation area and each segmentation area in the violation frame; selecting the maximum similarity as a region common factor corresponding to the target segmentation region;
the calculation formula of the similarity is as follows:
wherein,the similarity between the kth divided area in the characteristic frame of the mth frame and the kth divided area in the violation frame of the jth frame is obtained; x is X m,j The adaptive segmentation size is matched with the feature frame of the m frame and the corresponding violation frame of the j frame; gmk i The gray value of the ith pixel point in the kth partition area in the mth frame characteristic frame is obtained; gj' k i The gray value of the ith pixel point in the kth partition area in the jth frame violation frame is used as the gray value of the ith pixel point; gmk i′ The gray value of the ith pixel point in the kth partition area in the mth frame feature frame corresponds to the ith pixel point in the eighth adjacent area; gj' k i′ For the kth frame in the j-th frame violation frameThe ith pixel point in the partitioned area corresponds to the gray value of the ith' pixel point in the eighth adjacent area; exp is an exponential function with a natural constant as a base;
wherein the overall common factor is: an average value of region common factors of each divided region;
wherein, calculate the violation stability according to the fluctuation degree of whole public factor, include:
and for any feature frame, acquiring the integral common factors of all violation frames corresponding to the feature frame, and calculating the difference value of the integral common factors corresponding to adjacent violation frames, wherein the sum of the difference values is used as the violation stability.
2. The method for detecting 5G message content according to claim 1, wherein the calculation formula of the adaptive segmentation size is:
wherein X is m,j The adaptive segmentation size is matched with the feature frame of the m frame and the corresponding violation frame of the j frame; exp is an exponential function with a natural constant as a base; d (D) j The Hamming distance of the hash value of the m-th frame characteristic frame and the j-th frame illegal frame is used; AVG (D) m,J ) The average value of Hamming distances of hash values of the m-th frame characteristic frame and all corresponding offending frames is obtained; NG (NG) m The number of pixel points in the feature frame of the m-th frame; NG'. j The number of pixel points in the j-th frame violation frame is the number of pixel points in the j-th frame violation frame;to round the symbol up.
3. The method for detecting 5G message content according to claim 1, wherein said performing preliminary matching on the video information and the offending video in the offending video database to obtain a preliminary screening offending video includes:
acquiring a hash value of the video information as a first hash value; acquiring a hash value of any illegal video in the illegal video database as a second hash value; and calculating the Hamming distance between the first Hash value and the second Hash value, and selecting the illegal video with the Hamming distance smaller than a preset first threshold value as the primary screening illegal video.
4. The method for detecting 5G message content according to claim 1, wherein the filtering out the feature frames in the video image according to the information entropy of each frame of the video image in the video information comprises:
selecting any video image as a target image, acquiring information entropy corresponding to the target image, and calculating abnormal values of the target image relative to the front and rear frames of video images according to the information entropy corresponding to the target image;
and when the abnormal value is greater than or equal to a preset second threshold value, taking the target image as a characteristic frame in the video image.
5. The method for detecting 5G message content according to claim 4, wherein calculating an outlier of the target image with respect to the two video images according to the entropy of the information corresponding to the target image comprises:
acquiring information entropy of a front frame video image and a rear frame video image corresponding to a target image; calculating the average value of the information entropy of the target image and the front and rear two frames of video images, and taking the average value as the average value of the information entropy; calculating the absolute value of the difference value between the information entropy of the target image and the information entropy mean value; the ratio of the absolute value of the difference value and the information entropy of the target image is an abnormal value corresponding to the target image.
6. The method for detecting 5G message content according to claim 1, wherein the matching each frame of video image and feature frame in the preliminary screening violation video to obtain multi-frame violation frames includes:
acquiring a hash value of any characteristic frame as a third hash value; selecting any prescreened illegal video as a target video, and acquiring a hash value of any video image in the target video as a fourth hash value; calculating the Hamming distance of the third Hamming value and the fourth Hamming value, and taking a video image in the target video corresponding to the minimum Hamming distance as a violation frame;
and for any characteristic frame, obtaining a plurality of violation frames corresponding to the preliminary screening violation videos to obtain a plurality of frame violation frames.
7. The method for detecting content of a 5G message according to claim 1, wherein said identifying a violation of video information in the 5G message according to the violation stability comprises:
when the violation stability corresponding to any feature frame is greater than or equal to a preset third threshold, the corresponding video information is a violation video; otherwise, the corresponding video information is normal video.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211600092.8A CN115830508B (en) | 2022-12-13 | 2022-12-13 | 5G message content detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211600092.8A CN115830508B (en) | 2022-12-13 | 2022-12-13 | 5G message content detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115830508A CN115830508A (en) | 2023-03-21 |
| CN115830508B true CN115830508B (en) | 2024-03-08 |
Family
ID=85547012
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211600092.8A Active CN115830508B (en) | 2022-12-13 | 2022-12-13 | 5G message content detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115830508B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116866666B (en) * | 2023-09-05 | 2023-12-08 | 天津市北海通信技术有限公司 | Video stream picture processing method and device in rail transit environment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016082277A1 (en) * | 2014-11-28 | 2016-06-02 | 安科智慧城市技术（中国）有限公司 | Video authentication method and apparatus |
| CN109740424A (en) * | 2018-11-23 | 2019-05-10 | 深圳市华尊科技股份有限公司 | Traffic violations recognition methods and Related product |
| CN114581856A (en) * | 2022-05-05 | 2022-06-03 | 广东邦盛北斗科技股份公司 | Agricultural unit motion state identification method and system based on Beidou system and cloud platform |
-
2022
- 2022-12-13 CN CN202211600092.8A patent/CN115830508B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016082277A1 (en) * | 2014-11-28 | 2016-06-02 | 安科智慧城市技术（中国）有限公司 | Video authentication method and apparatus |
| CN109740424A (en) * | 2018-11-23 | 2019-05-10 | 深圳市华尊科技股份有限公司 | Traffic violations recognition methods and Related product |
| CN114581856A (en) * | 2022-05-05 | 2022-06-03 | 广东邦盛北斗科技股份公司 | Agricultural unit motion state identification method and system based on Beidou system and cloud platform |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115830508A (en) | 2023-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112200043B (en) | Intelligent danger source identification system and method for outdoor construction site | |
| US20230116801A1 (en) | Image authenticity detection method and device, computer device, and storage medium | |
| CN115830508B (en) | 5G message content detection method | |
| JP2012155727A (en) | Intra-mode region-of-interest image object segmentation | |
| CN108595422B (en) | Method for filtering bad multimedia messages | |
| CN111738120A (en) | Person identification method, person identification device, electronic equipment and storage medium | |
| CN112651319B (en) | Video detection method and device, electronic equipment and storage medium | |
| Zhang | Estimation of the mouth features using deformable templates | |
| CN114782892A (en) | Illegal behavior target detection method, device, equipment and storage medium | |
| CN115083008A (en) | Moving object detection method, device, equipment and storage medium | |
| CN110895811B (en) | Image tampering detection method and device | |
| CN113194281A (en) | Video analysis method and device, computer equipment and storage medium | |
| CN113792662B (en) | Image detection method, device, electronic equipment and storage medium | |
| CN113792661A (en) | Image detection method, image detection device, electronic equipment and storage medium | |
| EP1457927A2 (en) | Device and method for detecting blurring of image | |
| CN113822110A (en) | Target detection method and device | |
| CN112669244A (en) | Face image enhancement method and device, computer equipment and readable storage medium | |
| CN116364064A (en) | Audio splicing method, electronic equipment and storage medium | |
| CN114283431B (en) | Text detection method based on differentiable binarization | |
| CN115984973A (en) | Human body abnormal behavior monitoring method for peeping-proof screen | |
| CN102129575A (en) | Pornographic image analysis system based on color space skin color model | |
| CN114399497A (en) | Text image quality detection method and device, computer equipment and storage medium | |
| CN113014914A (en) | Neural network-based single face-changing short video identification method and system | |
| KR20200143041A (en) | Real-time detecting apparatus of video streaming error | |
| CN112906508A (en) | Face living body detection method based on convolutional neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |