CN111191055B - Method, device, computer equipment and storage medium for processing multimedia data - Google Patents
Method, device, computer equipment and storage medium for processing multimedia data Download PDFInfo
- Publication number
- CN111191055B CN111191055B CN202010002050.9A CN202010002050A CN111191055B CN 111191055 B CN111191055 B CN 111191055B CN 202010002050 A CN202010002050 A CN 202010002050A CN 111191055 B CN111191055 B CN 111191055B
- Authority
- CN
- China
- Prior art keywords
- attribute
- multimedia data
- multimedia
- region
- coefficient
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/43—Querying
- G06F16/435—Filtering based on additional data, e.g. user or group profiles
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention discloses a method, a device, a computer device and a storage medium for processing multimedia data, wherein the method comprises the following steps: determining multimedia data distributed in at least two regions; inquiring the multimedia attribute of the multimedia data; comparing multimedia data distributed in the same region as a whole to determine attribute weights of each multimedia attribute in each region; calculating a quality score representing quality for the multimedia data according to the multimedia attribute and the attribute weight calculation; and determining target multimedia data from the multimedia data in each region according to the quality scores. The personalized definition of the quality evaluation mode is realized, so that high-quality multimedia data is automatically screened, the workload of operators participating in screening of the multimedia data is greatly reduced, the cost of screening the high-quality multimedia data is greatly reduced, the efficiency of screening the high-quality multimedia data is improved, and the screening work of mass multimedia data in the Internet is adapted.
Description
Technical Field
Embodiments of the present invention relate to multimedia technologies, and in particular, to a method, an apparatus, a computer device, and a storage medium for processing multimedia data.
Background
With the rapid development of the internet, a large number of multimedia platforms are established on the internet, and the multimedia platforms provide a convenient way for users to share multimedia data.
These multimedia platforms preserve large amounts of multimedia data during re-operation and users upload multimedia data each day, potentially up to the millions of data volume level.
The multimedia data in the multimedia platform has different quality and huge data volume, and the user can select according to the keyword searching, subscribing and other modes by himself, so that the difficulty is high, and in order to provide services for the user, the multimedia platform generally selects some multimedia data with high quality to push to the user.
Because the different people's liveness and preference in each region are different, the quality evaluation modes are various, and especially, when live video data and the like are used in different regions, the requirement for screening multimedia data with higher quality by means of machine learning, deep learning and the like is difficult to reach.
Therefore, at present, the multimedia platform mostly introduces operators to manually screen multimedia data with higher quality aiming at different areas, and in the screening process, the operators need to manually intervene in the screening direction, but the cost of manual screening of the operators is higher, the efficiency is lower, and the screening work of mass multimedia data in the internet is more difficult to adapt.
Disclosure of Invention
The embodiment of the invention provides a processing method, a processing device, computer equipment and a storage medium for multimedia data, which are used for solving the problems of high requirements, high cost and low efficiency of manually screening multimedia data with high quality for different areas.
In a first aspect, an embodiment of the present invention provides a method for processing multimedia data, including:
determining multimedia data distributed in at least two regions;
inquiring the multimedia attribute of the multimedia data;
comparing multimedia data distributed in the same region as a whole to determine attribute weights of each multimedia attribute in each region;
calculating a quality score representing quality for the multimedia data according to the multimedia attribute and the attribute weight calculation;
and determining target multimedia data from the multimedia data in each region according to the quality scores.
In a second aspect, an embodiment of the present invention further provides a processing apparatus for multimedia data, including:
a multimedia data determining module for determining multimedia data distributed in at least two regions;
the multimedia attribute query module is used for querying the multimedia attribute of the multimedia data;
the attribute weight determining module is used for comparing the multimedia data distributed in the same area as a whole to determine the attribute weight of each multimedia attribute in each area;
the quality score calculation module is used for calculating the quality score of the characterization quality of the multimedia data according to the multimedia attribute and the attribute weight calculation;
and the target multimedia data determining module is used for determining target multimedia data from the multimedia data in each region according to the quality scores.
In a third aspect, an embodiment of the present invention further provides a computer apparatus, including:
one or more processors;
a memory for storing one or more programs;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of processing multimedia data as described in the first aspect.
In a fourth aspect, an embodiment of the present invention further provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor implements the method for processing multimedia data according to the first aspect.
In this embodiment, multimedia data published in at least two regions is determined, multimedia attributes of the multimedia data are queried, the multimedia data published in the same region are compared as a whole to determine attribute weights of each multimedia attribute in each region, mass fractions representing quality are calculated for the multimedia data according to the multimedia attributes and the attribute weights, target multimedia data are determined from the multimedia data in each region according to the mass fractions, the attribute weights adapted to the region are set for different regions in a unified manner, personalized definition of a quality evaluation manner is achieved, and therefore high-quality multimedia data are automatically screened, workload of operators participating in screening of the multimedia data is greatly reduced, cost of screening the high-quality multimedia data is greatly reduced, efficiency of screening the high-quality multimedia data is improved, and screening work of mass multimedia data in the internet is adapted.
Drawings
Fig. 1 is a flowchart of a method for processing multimedia data according to a first embodiment of the present invention;
fig. 2 is a flowchart of a method for processing multimedia data according to a second embodiment of the present invention;
fig. 3 is a schematic structural diagram of a multimedia data processing apparatus according to a third embodiment of the present invention;
fig. 4 is a schematic structural diagram of a computer device according to a fourth embodiment of the present invention.
Detailed Description
The invention is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting thereof. It should be further noted that, for convenience of description, only some, but not all of the structures related to the present invention are shown in the drawings.
Example 1
Fig. 1 is a flowchart of a method for processing multimedia data according to an embodiment of the present invention, where the method may be applied to determine attribute weights of multimedia attributes in multimedia data in different regions, so as to calculate quality scores of the multimedia data in different regions for filtering, and the method may be performed by a processing device of the multimedia data, where the processing device of the multimedia data may be implemented by software and/or hardware, and may be configured in a computer device, for example, a server, a workstation, a personal computer, etc., and the method specifically includes the following steps:
S101, determining multimedia data distributed in at least two areas.
In this embodiment, the multimedia platform operates across regions, and receives multimedia data uploaded by users in different regions.
In the same region, users accept the same or similar culture, interests and behaviors among users are similar, and quality evaluation is uniform, so that different multimedia data can be distinguished according to the released region.
For example, a region may refer to a country, such as china, thailand, vietnam, canada, etc., and of course, in order to improve accuracy in screening multimedia data, a region may refer to a fine-grained province, a local city, etc., which is not limited in this embodiment.
In addition, the types of multimedia data include video data, audio data, presentations, etc., and the forms include live programs, short videos, television shows, movies, electronic books, songs, etc.
Taking live programs as an example, the format of live programs is mostly streaming data, such as MP4 (Moving Picture Experts Group, fourth edition of moving picture experts group), FLV (Flash Video, streaming format), and the like.
The user logs in at the client, collects video data and audio data, encodes the video data and audio data, and transmits the encoded video data and audio data to the computer device through protocols such as RTSP (Real Time Streaming Protocol, real-time streaming protocol), RTMP (Real Time Messaging Protocol, real-time messaging protocol), HLS (HTTP Live Streaming, streaming media transmission protocol based on HTTP (Hyper Text Transport Protocol, hypertext transmission protocol)), and the like, and the computer device stores the live program when the live program is completed.
S102, inquiring the multimedia attribute of the multimedia data.
When multimedia data is generated and stored, the attribute is recorded as a multimedia attribute.
In a specific implementation, the multimedia attribute may be an attribute of the multimedia data itself, such as a title, an author (for a live program, the author is a host user), a length, etc., and the multimedia attribute may also be an attribute generated by interaction between the user and the multimedia data, such as positive emotion (like praise and support), negative emotion (like throwing eggs), comments, a bullet screen, a play amount, a viewing duration (a standard statistical duration/a length of the multimedia data, a standard statistical duration passes through a time detail of the user viewing the multimedia data, and an ETL (Extract-Transform-Load) in a hive table for describing a process of extracting (extracting), converting (transforming), and loading (Load) the data from a source end, which is generally represented by an average value of viewing durations of a plurality of users), a forwarding amount, and so on.
S103, comparing the multimedia data distributed in the same region as a whole to determine the attribute weight of each multimedia attribute in each region.
In this embodiment, the multimedia data distributed in the same region may be regarded as a whole, and the preference of the user in different regions is distinguished by comparing the differences of the multimedia data in different regions, so that the attribute weight is set for each multimedia attribute of the multimedia data in each region.
S104, calculating the quality score of the characterization quality of the multimedia data according to the multimedia attribute and the attribute weight.
In this embodiment, for each piece of multimedia data in each region, the multimedia attribute of the multimedia data and the attribute weight of the multimedia attribute in the region may be used, and substituted into a preset calculation formula, to generate a score for representing the quality of the multimedia data as a quality score.
S105, determining target multimedia data from the multimedia data in each region according to the quality scores.
If the multimedia score of the multimedia data in each region is calculated, the multimedia data with higher quality score can be selected from each region, that is, the multimedia data with better quality can be selected as the target multimedia data.
In one way of determining the target multimedia data, in each region, the multimedia data is sorted in descending order according to the quality score, that is, the larger the quality score is, the earlier the sorting is, the smaller the quality score is, and the later the sorting is, the first n pieces of multimedia data are determined to be the target multimedia data in the region, where n is a positive integer.
In this embodiment, multimedia data published in at least two regions is determined, multimedia attributes of the multimedia data are queried, the multimedia data published in the same region are compared as a whole to determine attribute weights of each multimedia attribute in each region, mass fractions representing quality are calculated for the multimedia data according to the multimedia attributes and the attribute weights, target multimedia data are determined from the multimedia data in each region according to the mass fractions, the attribute weights adapted to the region are set for different regions in a unified manner, personalized definition of a quality evaluation manner is achieved, and therefore high-quality multimedia data are automatically screened, workload of operators participating in screening of the multimedia data is greatly reduced, cost of screening the high-quality multimedia data is greatly reduced, efficiency of screening the high-quality multimedia data is improved, and screening work of mass multimedia data in the internet is adapted.
Example two
Fig. 2 is a flowchart of a processing method of multimedia data according to a second embodiment of the present invention, where the processing operations of determining attribute weights and calculating quality scores by collaborative filtering (Collaborative Filtering, abbreviated as CF) are further refined based on the foregoing embodiments, and the method specifically includes the following steps:
S201, determining multimedia data distributed in at least two regions.
S202, inquiring the multimedia attribute of the multimedia data.
S203, generating a first matrix.
In this embodiment, the first matrix includes attribute vectors that characterize each multimedia attribute for each region, i.e., the multimedia attributes in each region are converted into a form of vectors (i.e., attribute vectors).
In a specific implementation, for each region, an attribute ratio of each multimedia attribute in the region is determined, the attribute ratio being a ratio between the number of multimedia data having the multimedia attribute and the total number of multimedia data in the region.
For each region, multiplying the attribute duty ratio by an individual weight corresponding to the multimedia attribute to obtain an attribute vector, wherein the individual weight is a universal scoring weight which can be preset by a person skilled in the art according to service requirements.
For the length of the multimedia attribute, a preset plurality of ranges can be determined, each range is associated with a coefficient, the coefficient and the value of the range are in normal distribution, namely, the coefficient corresponding to a certain range in the middle is maximum, the coefficients corresponding to the ranges at two sides gradually decrease, at the moment, the range of the length of the multimedia video can be determined, the attribute duty ratio of the multimedia video in the range is counted, the coefficient associated with the range is set as the individual weight of the length, and the sum of the products between the attribute duty ratio corresponding to each range and the individual weight is calculated to be used as the attribute vector corresponding to the length.
And forming a first matrix by using the attribute vectors of at least two regions as rows and the multimedia attributes of the multimedia data as columns.
S204, generating a second matrix.
In this embodiment, the second matrix includes the similarity between every two regions on all the multimedia attributes, i.e., the similarity between all the multimedia attributes in each region is calculated.
In a specific implementation, for each two regions, the similarity between attribute vectors is calculated by means of cosine similarity, manhattan distance, hamming distance, and the like.
The similarity of at least two regions is formed into a second matrix with the regions as rows and the regions as columns.
S205, multiplying the second matrix by the first matrix to obtain the attribute weight of each item of multimedia attribute in each region.
In this embodiment, since the columns of the first matrix are regions and the behavior regions of the second matrix, the second matrix may be directly multiplied by the first matrix, and the product between the two is the attribute weight of each multimedia attribute in each region.
Let the second matrix a be an mxn matrix and the first matrix B be an nxp matrix, which are products AB being an mxp matrix.
(AB)[i,j]＝A[i,1]*B[1,j]+A[i,2]*B[2,j]+...+A[i,n]*B[n,j]
Wherein i ε m and j ε p.
In this embodiment, a first matrix is generated, where the first matrix includes attribute vectors representing each multimedia attribute in each region, a second matrix is generated, where the second matrix includes similarities between every two regions on all multimedia attributes, the second matrix is multiplied by the first matrix to obtain attribute weights of each multimedia attribute in each region, and accuracy of the attribute weights of each multimedia attribute in each region can be ensured by determining the attribute weights of each multimedia attribute in each region through the idea of collaborative filtering.
S206, converting the multimedia attribute into an attribute coefficient.
In this embodiment, the multimedia attribute is quantized by direct assignment, calculation of a numerical value with statistical significance, and the like, to obtain an attribute coefficient.
In one example, the multimedia attributes include at least one of:
title, forward emotion, author, play volume, viewing duration, length.
In this example, a title may be given a first coefficient, a second coefficient for positive emotion, a third coefficient for author, a fourth coefficient for length, as attribute coefficients.
The first coefficient, the second coefficient, the third coefficient and the fourth coefficient can be set by a person skilled in the art according to actual service requirements.
For example, if there is a header, one value (e.g., 1) is assigned as the first coefficient, and if there is no header, the other value (e.g., 0) is assigned as the first coefficient; if there is a positive emotion, one value (e.g., 1) is assigned as the first coefficient, and if there is no positive emotion, another value (e.g., 0) is assigned as the first coefficient; if there is an author, one value (e.g., 1) is assigned as the first coefficient, and if there is no author, another value (e.g., 0) is assigned as the first coefficient.
Further, when the fourth coefficient is given to the length, a plurality of preset ranges may be determined, each range is associated with a coefficient, the coefficient and the value of the range are normally distributed, that is, the coefficient corresponding to a certain range in the middle is maximum, the coefficients corresponding to the ranges at both sides gradually decrease, at this time, the range to which the length of the multimedia video belongs may be determined, and the coefficient associated with the range is set as the fourth coefficient of the length.
And carrying out dispersion standardization on the play quantity to obtain a first standard value serving as an attribute coefficient.
And carrying out dispersion standardization on the watching duration to obtain a second standard value as an attribute coefficient.
The influence of large units and small units (dimension elimination) and the influence of differences of variation sizes can be eliminated by the dispersion normalization, and the dispersion normalization can be calculated by the following formula:
X’＝(X-X min )/(X max -X min )
Wherein X is play amount or viewing duration, X max X is the maximum value of play quantity or viewing duration min For maximum play or viewing duration, X' is the first standard value or the second standard value.
Of course, the attribute coefficients and the conversion manners thereof are merely examples, and other attribute coefficients and conversion manners thereof may be set according to actual situations when the present embodiment is implemented, and the present embodiment is not limited thereto. In addition, other attribute coefficients and transformation modes thereof can be adopted by those skilled in the art according to actual needs, and the embodiment is not limited thereto.
S207, calculating the original score of the multimedia data by using the attribute coefficient and the attribute weight.
After determining the attribute coefficients, then, an original score for the multimedia data, which characterizes the quality of the multimedia data basis, may be calculated in combination with the attribute coefficients and the attribute weights.
The original score is positively correlated with the attribute coefficient and the attribute weight, namely, the larger the attribute coefficient is, the higher the attribute weight is, and the higher the original score is, otherwise, the smaller the attribute coefficient is, the lower the attribute weight is, and the lower the original score is.
In one example, a first product between the first coefficient and the attribute weight corresponding to the title is calculated, a second product between the second coefficient and the attribute weight corresponding to the forward emotion is calculated, a third product between the third coefficient and the attribute weight corresponding to the author is calculated, a fourth product between the first standard value and the attribute weight corresponding to the play amount is calculated, and a fifth product between the second standard value and the attribute weight corresponding to the viewing duration is calculated.
A sum is calculated between the first product, the second product, the third product, the fourth product, and the fifth product.
A sixth product between the sum and the fourth coefficient is calculated as an original score of the multimedia data.
Of course, the above-mentioned calculation method of the original score is merely an example, and in implementing the present embodiment, other calculation methods of the original score may be set according to the actual situation of the attribute coefficient and the attribute weight, for example, products between all attribute coefficients and all attribute weights are calculated, and then summed, as the original score, etc., which is not limited in this embodiment. In addition, in addition to the above-mentioned calculation method of the original score, those skilled in the art may also use other calculation methods of the original score according to actual needs, which is not limited in this embodiment.
And S208, attenuating the original scores to obtain candidate scores.
In this embodiment, for multimedia data with the same content, quality changes may be caused by different factors, so that the original score may be attenuated according to the factors, and the original score may be reduced as a candidate score.
Further, an example in which time is taken as a factor is illustrated, in which case the distribution time of multimedia data may be determined.
And determining an attenuation coefficient which is larger than 0 and smaller than 1 and is inversely related to the time difference value, wherein the time difference value is the difference value between the release time and the current time, and the larger the time difference value is, the smaller the attenuation coefficient is, and conversely, the smaller the time difference value is, the larger the attenuation coefficient is.
In one example, newton's law of cooling may be employed, issuing a time decay factor, i.e. setting an exponential function as the decay formula: y=a x Where y is the decay factor, x is the time difference, a is a constant determined based on half-life, and if half-life is 30, a≡0.977.
At this time, a seventh product between the original score and the attenuation coefficient may be calculated as a candidate score.
Of course, the foregoing manner of attenuation is merely an example, and other manners of attenuation may be set according to actual situations when implementing the present embodiment, for example, for a plurality of multimedia data with the same or similar content, attenuation coefficients may be determined according to at least one of the forwarding times, the praying amounts, and the bullet screen amounts, where the forwarding times, the praying amounts, the bullet screen amounts, and the attenuation coefficients are all positively correlated, so as to reduce pushing of the multimedia data with the same or similar content to the user, and so on, and the embodiment of the present invention is not limited thereto. In addition, in addition to the above-described judgment processing method, a person skilled in the art may adopt other judgment processing methods according to actual needs, which is not limited in this embodiment.
And S209, adding a preset basic score on the basis of the candidate score to serve as a quality score of the multimedia data representation quality.
In this embodiment, an operator may set a basic score for multimedia data in different regions in advance according to dimensions such as types, authors, and subjects of the multimedia data, so that the operator may manually guide screening of the multimedia data through human intervention.
For example, in a global large electronic competition, an operator may set a higher base score for a topic of the electronic competition, thereby increasing the push of live programs for the topic of the electronic competition.
For example, for a live program or a short video, the multimedia platform newly signs up for an excellent anchor user or studio, and an operator can set a higher base score with the anchor user or studio as an author, so as to increase the push amount of the live program or the short video made by the anchor user in the initial stage.
For another example, in a global holiday, such as a tree planting festival of 3 months and 12 days, a fire-fighting propaganda day of 11 months and 9 days, etc., an operator may set a higher base score for a live program or short video on which the holiday is the subject, thereby increasing the push amount of the live program or short video on which the holiday is the subject.
In this embodiment, the multimedia attribute is converted into the attribute coefficient, the attribute coefficient and the attribute weight are used to calculate the original score for the multimedia data, the original score is positively correlated with the attribute coefficient and the attribute weight, the original score is attenuated to obtain the candidate score, the preset basic score is added on the basis of the candidate score to be used as the quality score of the multimedia data characterization quality, the multimedia attribute represents the original quality of the multimedia data, on the basis of the original score, the attenuation is carried out according to different factors, and the basic score which can be allocated by operators is added, so that the adjustment of the original score in different dimensionalities is realized, and the accuracy of the quality characterization quality of the quality score is ensured.
And S210, determining target multimedia data from the multimedia data in each region according to the quality scores.
S211, pushing the target multimedia data in each region to users in the region.
In this embodiment, for each region, the target multimedia data screened out may be directly pushed to the user in the region, so as to provide the multimedia data with higher quality suitable for the region to the user.
Of course, in addition to direct pushing, the target multimedia data may be screened again through other dimensions, for example, matching the target multimedia data with a tag (such as preference, age, gender, etc.) characterizing the individuation of the user, and the multimedia data after being screened again is pushed to the user, which is not limited in this embodiment.
It should be noted that, the user is characterized by a user identifier, and when the user is registered, the user identifier may be a user ID, a user account, or the like, and when the user is not registered, the user identifier may be a device identifier, or the like, so that the user is pushed to the device where the user identifier is located for display, such as a mobile terminal, a personal computer, or the like.
In order to enable those skilled in the art to better understand the present embodiment, a method for processing multimedia data in the present embodiment is described below by way of specific examples.
IN this example, the regions include thailand TH, vietnam VN, brazil BR, japan JA, indian IN, canada CA.
The multimedia attributes include title, forward emotion, author, play volume, viewing duration, length.
The coefficients associated with the range of 0-30 seconds are 1, the coefficients associated with the range of 31-60 seconds are 3, and the coefficients associated with the range of 61 seconds or more are 2.
The title has an individual weight of 3, a forward emotion has an individual weight of 5, an author has an individual weight of 1, a play amount has an individual weight of 100, a viewing time length has an individual weight of 200, a length of 0-30 seconds has an individual weight of 1, a length of 31-60 seconds has an individual weight of 3, and a length of 61 seconds or more has an individual weight of 2.
For the convenience of calculation, for the play amount, statistics may be performed using a first standard value obtained by performing dispersion normalization on the play amount, and for the viewing duration, statistics may be performed using a second standard value obtained by performing dispersion normalization on the viewing duration.
The attribute ratios of the multimedia data in the areas are counted and attribute vectors are calculated, and the following table is shown:
the three attribute ratios corresponding to the length are sequentially an attribute ratio of 0-30 seconds, an attribute ratio of 31-60 seconds and an attribute ratio of more than 61 seconds.
The first matrix is as follows:
| P1 | P2 | P3 | P4 | P5 | P6 | |
| TH | 0.3 | 0.75 | 0.12 | 6 | 4 | 2.6 |
| VN | 0.06 | 1.25 | 0.13 | 16 | 44 | 4.76 |
| BR | 0.306 | 0.55 | 0.13 | 60 | 42 | 5.2 |
| JA | 0.27 | 0.075 | 0.19 | 6.5 | 5.2 | 2.03 |
| IN | 0.06 | 0.15 | 0.045 | 6.1 | 2.4 | 0.834 |
| CA | 0.33 | 0.75 | 0.125 | 6 | 5 | 3.56 |
calculating the similarity of multimedia attributes between every two regions through cosine similarity, taking TH and VN as examples:
the second matrix is as follows:
multiplying the second matrix by the first matrix to obtain attribute weights of each region as follows:
| P1 | P2 | P3 | P4 | P5 | P6 | |
| TH | 1.29 | 3.22 | 0.70 | 94.41 | 91.46 | 17.68 |
| VN | 1.09 | 3.07 | 0.63 | 84.47 | 91.73 | 16.28 |
| BR | 1.27 | 3.21 | 0.70 | 96.75 | 94.00 | 17.71 |
| JA | 1.30 | 3.30 | 0.71 | 96.52 | 94.90 | 18.03 |
| IN | 1.24 | 2.99 | 0.67 | 92.36 | 85.91 | 16.71 |
| CA | 1.28 | 3.24 | 0.70 | 92.78 | 91.38 | 17.62 |
assuming that there is one multimedia data in TH, there is title, forward emotion, no author, a first standard value for dispersion normalization of play amount is 0.012, a second standard value for dispersion normalization of viewing duty is 0.045, the length of the multimedia data is 35 seconds, and the multimedia data has been distributed for 30 days.
Title 1 is given as a first coefficient, positive emotion 1 is given as a second coefficient, author 0 is given as a third coefficient, and length 35 description is given as a fourth coefficient by a coefficient 3 associated with a range of 31-60 seconds.
Substituting 30 days as x into the decay formula y=a x Wherein the half-life is 30, the attenuation coefficient y=0.5.
The base score set by the operator for the multimedia data is 10000.
Therefore, the mass fraction of the multimedia data is:
(1×1.29+1×3.22+0×0.70+94.41×0.012+91.46×0.045)×3×17.68×0.5+10000＝10258.8
example III
Fig. 3 is a schematic structural diagram of a multimedia data processing apparatus according to a third embodiment of the present invention, where the apparatus may specifically include the following modules:
a multimedia data determining module 301 for determining multimedia data distributed in at least two regions;
a multimedia attribute query module 302, configured to query multimedia attributes of the multimedia data;
an attribute weight determining module 303, configured to compare multimedia data distributed in the same region as a whole, so as to determine an attribute weight of each multimedia attribute in each region;
a quality score calculating module 304, configured to calculate a quality score representing quality for the multimedia data according to the multimedia attribute and the attribute weight calculation;
a target multimedia data determining module 305 for determining target multimedia data from the multimedia data in each region according to the quality score.
In one embodiment of the present invention, the attribute weight determining module 303 includes:
A first matrix generation sub-module, configured to generate a first matrix, where the first matrix includes an attribute vector that characterizes each multimedia attribute in each region;
a second matrix generation sub-module, configured to generate a second matrix, where the second matrix includes similarities between every two regions on all multimedia attributes;
and the matrix phase multiplication sub-module is used for multiplying the second matrix by the first matrix to obtain the attribute weight of each item of multimedia attribute in each region.
In one embodiment of the present invention, the first matrix generation submodule includes:
an attribute duty ratio determining unit configured to determine, for each region, an attribute duty ratio of each multimedia attribute in the region;
the attribute vector determining unit is used for multiplying the attribute duty ratio by the individual weight corresponding to the multimedia attribute for each region to obtain an attribute vector;
and the first matrix composition unit is used for composing the attribute vectors of at least two areas into a first matrix.
In one embodiment of the present invention, the second matrix generation submodule includes:
a similarity calculation unit configured to calculate, for each two regions, a similarity between the attribute vectors;
And the second matrix composing unit is used for composing the similarity of at least two areas into a second matrix.
In one embodiment of the present invention, the mass fraction calculation module 304 includes:
the attribute coefficient conversion sub-module is used for converting the multimedia attribute into an attribute coefficient;
an original score calculation sub-module, configured to calculate an original score for the multimedia data using the attribute coefficient and the attribute weight, where the original score is positively correlated with both the attribute coefficient and the attribute weight;
the original score attenuation submodule is used for attenuating the original score to obtain a candidate score;
and the basic score adding sub-module is used for adding a preset basic score on the basis of the candidate score to be used as a quality score of the multimedia data representation quality.
In one example of an embodiment of the present invention, the multimedia attribute includes at least one of:
title, forward emotion, author, play volume, viewing duration, length;
the attribute coefficient conversion submodule includes:
a first conversion unit configured to assign a first coefficient to the title, a second coefficient to the positive emotion, a third coefficient to the author, and a fourth coefficient to the length as attribute coefficients;
The second conversion unit is used for carrying out dispersion standardization on the play quantity to obtain a first standard value which is used as an attribute coefficient;
and the third conversion unit is used for carrying out dispersion normalization on the watching duration to obtain a second standard value as an attribute coefficient.
In this example, the raw score computation submodule includes:
a first product calculation unit, configured to calculate a first product between the first coefficient and an attribute weight corresponding to the title;
a second product calculation unit, configured to calculate a second product between the second coefficient and the attribute weight corresponding to the forward emotion;
a third product calculation unit, configured to calculate a third product between the third coefficient and the attribute weight corresponding to the author;
a fourth product calculation unit, configured to calculate a fourth product between the first standard value and the attribute weight corresponding to the play amount;
a fifth product calculating unit, configured to calculate a fifth product between the second standard value and the attribute weight corresponding to the viewing duration;
a sum value calculation unit configured to calculate a sum value between the first product, the second product, the third product, the fourth product, and the fifth product;
A sixth product calculation unit for calculating a sixth product between the sum and the fourth coefficient as an original score of the multimedia data.
In one embodiment of the invention, the raw score attenuation submodule includes:
a release time determining unit, configured to determine a release time of the multimedia data;
the attenuation coefficient determining unit is used for determining an attenuation coefficient which is smaller than 1 and is inversely related to a time difference value, wherein the time difference value is a difference value between the release time and the current time;
a seventh product calculation unit, configured to calculate a seventh product between the original score and the attenuation coefficient as a candidate score.
In one embodiment of the present invention, the target multimedia data determining module 305 includes:
a descending order sorting sub-module, configured to sort the multimedia data in descending order according to the quality score in each region;
and the sequencing determination submodule is used for determining the first n pieces of multimedia data sequenced as target multimedia data in the region.
In one embodiment of the present invention, further comprising:
and the target multimedia data pushing module is used for pushing the target multimedia data in each region to users in the region.
The multimedia data processing device provided by the embodiment of the invention can execute the multimedia data processing method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Example IV
Fig. 4 is a schematic structural diagram of a computer device according to a fourth embodiment of the present invention. As shown in fig. 4, the computer apparatus includes a processor 400, a memory 401, a communication module 402, an input device 403, and an output device 404; the number of processors 400 in the computer device may be one or more, one processor 400 being taken as an example in fig. 4; the processor 400, the memory 401, the communication module 402, the input means 403 and the output means 404 in the computer device may be connected by a bus or in other ways, in fig. 4 by way of example.
The memory 401 is a computer-readable storage medium, and may be used to store a software program, a computer-executable program, and modules corresponding to a processing method of multimedia data in this embodiment (for example, a multimedia data determination module 301, a multimedia attribute query module 302, an attribute weight determination module 303, a quality score calculation module 304, and a target multimedia data determination module 305 in a processing apparatus of multimedia data as shown in fig. 3). The processor 400 executes various functional applications of the computer device and data processing, i.e., implements the above-described multimedia data processing method, by running software programs, instructions, and modules stored in the memory 401.
The memory 401 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, at least one application program required for functions; the storage data area may store data created according to the use of the computer device, etc. In addition, memory 401 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage device. In some examples, memory 401 may further include memory located remotely from processor 400, which may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
And the communication module 402 is used for establishing connection with the display screen and realizing data interaction with the display screen.
The input means 403 may be used for receiving input numeric or character information and generating key signal inputs related to user settings and function control of the computer device, as well as a camera for capturing images and a sound pickup device for capturing audio data.
The output 404 may include an audio device such as a speaker.
The specific composition of the input device 403 and the output device 404 may be set according to the actual situation.
The processor 400 executes various functional applications of the device and data processing by running software programs, instructions and modules stored in the memory 401, i.e., implements the above-described connection node control method of the electronic whiteboard.
The computer device provided in this embodiment may execute the multimedia data processing method provided in any one of the embodiments of the present invention, and specifically, the corresponding functions and beneficial effects.
Example five
A fifth embodiment of the present invention also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method for processing multimedia data, the method comprising:
determining multimedia data distributed in at least two regions;
inquiring the multimedia attribute of the multimedia data;
comparing multimedia data distributed in the same region as a whole to determine attribute weights of each multimedia attribute in each region;
calculating a quality score representing quality for the multimedia data according to the multimedia attribute and the attribute weight calculation;
And determining target multimedia data from the multimedia data in each region according to the quality scores.
Of course, the computer readable storage medium provided by the embodiments of the present invention, the computer program thereof is not limited to the method operations described above, and related operations in the multimedia data processing method provided by any embodiment of the present invention may also be performed.
From the above description of embodiments, it will be clear to a person skilled in the art that the present invention may be implemented by means of software and necessary general purpose hardware, but of course also by means of hardware, although in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, etc., and include several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments of the present invention.
It should be noted that, in the embodiment of the processing apparatus for multimedia data, each unit and module included are only divided according to the functional logic, but not limited to the above-mentioned division, so long as the corresponding functions can be implemented; in addition, the specific names of the functional units are also only for distinguishing from each other, and are not used to limit the protection scope of the present invention.
Note that the above is only a preferred embodiment of the present invention and the technical principle applied. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, while the invention has been described in connection with the above embodiments, the invention is not limited to the embodiments, but may be embodied in many other equivalent forms without departing from the spirit or scope of the invention, which is set forth in the following claims.
Claims (11)
1. A method for processing multimedia data, comprising:
determining multimedia data distributed in at least two regions;
Inquiring the multimedia attribute of the multimedia data;
comparing multimedia data distributed in the same region as a whole to determine attribute weights of each multimedia attribute in each region;
calculating a quality score representing quality for the multimedia data according to the multimedia attribute and the attribute weight;
determining target multimedia data from the multimedia data in each region according to the quality score;
the comparing the multimedia data distributed in the same region as a whole to determine the attribute weight of each multimedia attribute in each region comprises the following steps:
generating a first matrix, wherein the first matrix comprises attribute vectors for each region to represent each item of multimedia attribute;
generating a second matrix, wherein the second matrix comprises the similarity between every two areas on all multimedia attributes;
and multiplying the second matrix by the first matrix to obtain the attribute weight of each multimedia attribute in each region.
2. The method of claim 1, wherein the generating a first matrix comprises:
determining, for each region, an attribute duty cycle of each multimedia attribute in the region;
Multiplying the attribute duty ratio by the individual weight corresponding to the multimedia attribute for each region to obtain an attribute vector;
and forming a first matrix by the attribute vectors of at least two areas.
3. The method according to claim 1 or 2, wherein the generating a second matrix comprises:
calculating the similarity between the attribute vectors for every two regions;
and forming a second matrix by the similarity of at least two areas.
4. The method according to claim 1 or 2, wherein said calculating a quality score characterizing quality for said multimedia data from said multimedia attributes and said attribute weight calculation comprises:
converting the multimedia attribute into an attribute coefficient;
calculating an original score for the multimedia data by using the attribute coefficient and the attribute weight, wherein the original score is positively correlated with the attribute coefficient and the attribute weight;
attenuating the original score to obtain a candidate score;
and adding a preset basic score on the basis of the candidate score to serve as a quality score of the multimedia data representation quality.
5. The method of claim 4, wherein the multimedia attributes comprise at least one of:
Title, forward emotion, author, play volume, viewing duration, length;
the converting the multimedia attribute into an attribute coefficient comprises:
assigning a first coefficient to the title, a second coefficient to the positive emotion, a third coefficient to the author, and a fourth coefficient to the length as attribute coefficients;
performing dispersion standardization on the play quantity to obtain a first standard value serving as an attribute coefficient;
performing dispersion standardization on the watching duration to obtain a second standard value serving as an attribute coefficient;
the computing an original score for the multimedia data using the attribute coefficients and the attribute weights comprises:
calculating a first product between the first coefficient and the attribute weight corresponding to the title;
calculating a second product between the second coefficient and the attribute weight corresponding to the forward emotion;
calculating a third product between the third coefficient and the attribute weight corresponding to the author;
calculating a fourth product between the first standard value and the attribute weight corresponding to the play quantity;
calculating a fifth product between the second standard value and the attribute weight corresponding to the viewing duration;
Calculating a sum between the first product, the second product, the third product, the fourth product, and the fifth product;
a sixth product between the sum and the fourth coefficient is calculated as an original score of the multimedia data.
6. The method of claim 4, wherein attenuating the original score to obtain a candidate score comprises:
determining the release time of the multimedia data;
determining an attenuation coefficient which is smaller than 1 and is inversely related to a time difference value, wherein the time difference value is the difference value between the release time and the current time;
a seventh product between the original score and the decay factor is calculated as a candidate score.
7. The method according to claim 1 or 2 or 5 or 6, wherein said determining target multimedia data from multimedia data in each region according to said quality score comprises:
in each region, ordering the multimedia data in descending order according to the quality score;
and determining the first n multimedia data of the sequence as target multimedia data in the region.
8. The method of claim 1 or 2 or 5 or 6, further comprising:
The target multimedia data in each region is pushed to users in the region.
9. A processing apparatus for multimedia data, comprising:
a multimedia data determining module for determining multimedia data distributed in at least two regions;
the multimedia attribute query module is used for querying the multimedia attribute of the multimedia data;
the attribute weight determining module is used for comparing the multimedia data distributed in the same area as a whole to determine the attribute weight of each multimedia attribute in each area;
the quality score calculating module is used for calculating the quality score of the characterization quality for the multimedia data according to the multimedia attribute and the attribute weight;
a target multimedia data determining module for determining target multimedia data from the multimedia data in each region according to the quality score;
the attribute weight determination module includes:
a first matrix generation sub-module, configured to generate a first matrix, where the first matrix includes an attribute vector that characterizes each multimedia attribute in each region;
a second matrix generation sub-module, configured to generate a second matrix, where the second matrix includes similarities between every two regions on all multimedia attributes;
And the matrix phase multiplication sub-module is used for multiplying the second matrix by the first matrix to obtain the attribute weight of each item of multimedia attribute in each region.
10. A computer device, the computer device comprising:
one or more processors;
a memory for storing one or more programs;
when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the method of processing multimedia data as claimed in any one of claims 1-8.
11. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements a method of processing multimedia data according to any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010002050.9A CN111191055B (en) | 2020-01-02 | 2020-01-02 | Method, device, computer equipment and storage medium for processing multimedia data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010002050.9A CN111191055B (en) | 2020-01-02 | 2020-01-02 | Method, device, computer equipment and storage medium for processing multimedia data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111191055A CN111191055A (en) | 2020-05-22 |
| CN111191055B true CN111191055B (en) | 2023-06-16 |
Family
ID=70709662
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010002050.9A Active CN111191055B (en) | 2020-01-02 | 2020-01-02 | Method, device, computer equipment and storage medium for processing multimedia data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111191055B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111782831A (en) * | 2020-07-02 | 2020-10-16 | 北京字节跳动网络技术有限公司 | Media file display method, device, medium and electronic equipment |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140344861A1 (en) * | 2013-05-14 | 2014-11-20 | Tivo Inc. | Method and system for trending media programs for a user |
| CN105120345B (en) * | 2015-08-20 | 2018-08-14 | 广州创维平面显示科技有限公司 | A kind of method for ordering television channels and system based on range statistics and switching times |
| CN105868237A (en) * | 2015-12-09 | 2016-08-17 | 乐视网信息技术（北京）股份有限公司 | Multimedia data recommendation method and server |
| CN110598016B (en) * | 2019-09-11 | 2021-08-17 | 腾讯科技（深圳）有限公司 | Method, device, equipment and medium for recommending multimedia information |
-
2020
- 2020-01-02 CN CN202010002050.9A patent/CN111191055B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111191055A (en) | 2020-05-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109408639B (en) | Bullet screen classification method, bullet screen classification device, bullet screen classification equipment and storage medium | |
| CN105138541B (en) | The method and apparatus of audio-frequency fingerprint matching inquiry | |
| WO2018090793A1 (en) | Multimedia recommendation method and device | |
| US9805022B2 (en) | Generation of topic-based language models for an app search engine | |
| CN106326391B (en) | Multimedia resource recommendation method and device | |
| CN109862397B (en) | Video analysis method, device, equipment and storage medium | |
| US20170195753A1 (en) | System and method for generating segmented content based on related data ranking | |
| US20150112814A1 (en) | System and method for an integrated content publishing system | |
| WO2013184957A1 (en) | Systems and methods of classifying content items | |
| CN109429103B (en) | Method and device for recommending information, computer readable storage medium and terminal equipment | |
| CN107977678B (en) | Method and apparatus for outputting information | |
| CN106303556A (en) | Video resource call method, Apparatus and system | |
| CN111191055B (en) | Method, device, computer equipment and storage medium for processing multimedia data | |
| CN111259245B (en) | Work pushing method, device and storage medium | |
| CN112188295A (en) | Video recommendation method and device | |
| WO2015063055A1 (en) | Method and system for providing access to auxiliary information | |
| CN104965903A (en) | Resource recommendation method and apparatus | |
| CN114491149A (en) | Information processing method and apparatus, electronic device, storage medium, and program product | |
| JP2018088051A (en) | Information processing device, information processing method and program | |
| CN111104583A (en) | Live broadcast room recommendation method, storage medium, electronic device and system | |
| CN107124660B (en) | Method and device for triggering broadcast information of live broadcast room | |
| CN104657383B (en) | A kind of repetition video detecting method and system based on associate feature | |
| US20150278352A1 (en) | Systems and methods for sharing audio feeds | |
| CN105141982A (en) | Method and device for generating electronic program guide (EPG) of hot programs | |
| RU2530671C1 (en) | Checking method of web pages for content in them of target audio and/or video (av) content of real time |
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 |