CN107153809B - Method and device for confirming television station icon - Google Patents
Method and device for confirming television station icon Download PDFInfo
- Publication number
- CN107153809B CN107153809B CN201610124860.5A CN201610124860A CN107153809B CN 107153809 B CN107153809 B CN 107153809B CN 201610124860 A CN201610124860 A CN 201610124860A CN 107153809 B CN107153809 B CN 107153809B
- Authority
- CN
- China
- Prior art keywords
- similar
- television station
- station
- probability
- icon
- 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.)
- Expired - Fee Related
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
- User Interface Of Digital Computer (AREA)
Abstract
The invention discloses a method and a device for confirming television station icons. The method for confirming the television station icon comprises the following steps: acquiring n frame images of a television station to be detected; respectively analyzing similar television station icons m before the similar probability ranking of each frame image in the n frames of images of the television station to be detected, wherein the similar television station icons comprise similar probabilities; discarding the similarity probability of an anomaly; after discarding the abnormal similar probability, counting the cumulative similar probability sum of each similar TV station icon; and determining the accumulated similarity probability and the highest similar television station icon as a matched television station icon. The invention can accurately identify the television station icon in the picture of the television station to be identified so as to acquire the information related to the television station icon in the following.
Description
Technical Field
The invention relates to the field of image recognition, in particular to a method and a device for confirming television station icons.
Background
The television station icon is one of the special visual contents in the broadcast video, contains important semantic information of the television station such as the station name, the type, the copyright and the like, is an important identifier for distinguishing the broadcast television channels, and the identification of the station icon has important significance for program guide, content analysis, retrieval and the like of the broadcast television. Currently, in the identification of station icons, it is often encountered that station icons appear in complex background images, resulting in the appearance of extreme conditions, known as "image noise" in some cases. Especially in the case of a small number of samples, individual "image noise" can seriously affect the accuracy of identifying the tv station icon. At present, the number of television stations is very large, at least three hundred channels are counted, the similarity of some television station icons is higher, and the difficulty of identifying the television station icons is increased. Therefore, how to solve the above problems is an urgent issue to be solved in the industry.
Disclosure of Invention
The invention provides a method and a device for confirming a television station icon, which are used for improving the accuracy of confirming the television station icon.
According to a first aspect of the embodiments of the present invention, there is provided a method for confirming a tv station icon, including:
acquiring n frame images of a television station to be detected;
respectively analyzing similar television station icons m before the similar probability ranking of each frame image in the n frames of images of the television station to be detected, wherein the similar television station icons comprise similar probabilities;
discarding the similarity probability of an anomaly;
after discarding the abnormal similar probability, counting the cumulative similar probability sum of each similar TV station icon;
and determining the accumulated similarity probability and the highest similar television station icon as a matched television station icon.
In one embodiment, the analyzing step of respectively analyzing similar tv station icons ranked m before the similar probability of each of the n frames of images of the tv station to be detected, where the similar tv station icons include similar probabilities, includes:
analyzing similar TV station icons of the ith frame image of the TV station to be detected, wherein the similar TV station icons comprise similar probability;
confirming that the television station icon m before the similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon;
and traversing each frame image in the n frames of images of the television station to be detected, and analyzing the similar television station icons m before the similar probability ranking of each frame image in the n frames of images of the television station to be detected.
In one embodiment, said discarding said similarity probabilities of anomalies comprises:
counting the cumulative similarity probability sum of the jth similar television station icon;
calculating the average similarity probability of the jth similar television station icon;
calculating the variance of each similarity probability of the jth similar television station icon and the average similarity probability of the jth similar television station icon;
judging whether the variance is larger than a preset variance threshold value or not;
when the variance is larger than a preset variance threshold value, confirming that the similarity probability corresponding to the variance is the abnormal similarity probability;
discarding the likelihood probability of the anomaly;
traversing all of the similar TV station icons, and discarding the abnormal similar probabilities of all of the similar TV station icons.
In one embodiment, said counting the cumulative sum of similarity probabilities for each similar tv station icon after discarding said similarity probabilities for said anomalies comprises:
after the abnormal similar probability is abandoned, the cumulative similar probability sum of the jth similar TV station icon is counted again;
and traversing all the icons of the similar television stations, and respectively counting the cumulative similarity probability sums of all the similar television stations.
In one embodiment, the confirming that the tv station icon m before the similarity probability ranking of the ith frame image of the tv station to be detected is a similar tv station icon includes:
judging whether the similarity probabilities m before the similarity probability ranking of the ith frame image of the television station to be detected are all larger than or equal to a preset limited probability threshold value;
when the judgment is yes, confirming that the television station icon m before the similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon;
if not, comparing the similar television station icons with the similarity probability of the ith frame image of the television station to be detected being greater than or equal to the similarity probability ranking k before the preset limited probability threshold, wherein k is smaller than m;
calculating the similar television station icon of the k before the similarity probability ranking of the ith frame image of the television station to be detected;
and when the k is equal to 0, determining that the ith frame image of the television station to be detected does not have a corresponding similar television station icon.
According to a second aspect of the embodiments of the present invention, there is provided an apparatus for confirming a tv station icon, including:
the acquisition module is used for acquiring n frame images of a television station to be detected;
the analysis module is used for respectively analyzing similar television station icons m before the similar probability ranking of each frame image in the n frames of images of the television station to be detected, wherein the similar television station icons comprise similar probabilities;
a discarding module for discarding the similarity probability of an anomaly;
the statistic module is used for counting the cumulative similarity probability sum of each similar TV station icon after discarding the abnormal similarity probability;
and the determining module is used for determining the accumulated similarity probability and the highest similar TV station icon as a matched TV station icon.
In one embodiment, the analysis module comprises:
the analysis submodule is used for analyzing similar television station icons of the ith frame image of the television station to be detected, and the similar television station icons comprise similar probabilities;
the first confirming submodule is used for confirming that the television station icon m before the similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon;
the first confirming submodule is also used for judging whether the similarity probabilities m before the similarity probability ranking of the ith frame image of the television station to be detected are all larger than or equal to a preset limited probability threshold value; when the judgment is yes, confirming that the television station icon m before the similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon; if not, comparing the similar television station icons with the similarity probability of the ith frame image of the television station to be detected being greater than or equal to the similarity probability ranking k before the preset limited probability threshold, wherein k is smaller than m; calculating the similar television station icon of the k before the similarity probability ranking of the ith frame image of the television station to be detected; and when the k is equal to 0, determining that the ith frame image of the television station to be detected does not have a corresponding similar television station icon.
And the first traversal submodule is used for traversing each frame image in the n frame images of the television station to be detected and analyzing the similar television station icon m before the similar probability ranking of each frame image in the n frame images of the television station to be detected.
In one embodiment, the discard module comprises:
the first statistic submodule is used for counting the cumulative similarity probability sum of the jth similar television station icon;
the first calculation submodule is used for calculating the average similarity probability of the jth similar television station icon;
the second calculation submodule is used for calculating the variance between each similarity probability of the jth similar television station icon and the average similarity probability of the jth similar television station icon;
the judgment submodule is used for judging whether the variance is larger than a preset variance threshold value or not;
the second confirming submodule is used for confirming that the similarity probability corresponding to the variance is abnormal when the variance is larger than a preset variance threshold;
a discard submodule for discarding the likelihood probability of the anomaly;
and the second traversal submodule is used for traversing all the similar television station icons and abandoning the abnormal similar probability of all the similar television station icons.
In one embodiment, the statistics module includes:
the second statistic submodule is used for counting the cumulative similarity probability sum of the jth similar TV station icon again after the abnormal similarity probability is discarded;
and the third traversal submodule is used for traversing all the icons of the similar television stations and respectively counting the cumulative similarity probability sum of all the similar television stations.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart illustrating a method of confirming a television station icon in accordance with an exemplary embodiment of the present invention;
fig. 2 is a flowchart illustrating a step S12 of a method for confirming a tv station icon according to an exemplary embodiment of the present invention;
fig. 3 is a flowchart illustrating a step S13 of a method for confirming a tv station icon according to an exemplary embodiment of the present invention;
fig. 4 is a flowchart illustrating a step S14 of a method for confirming a tv station icon according to an exemplary embodiment of the present invention;
fig. 5 is a flowchart illustrating a step S22 of a method for confirming a tv station icon according to another exemplary embodiment of the present invention;
fig. 6 is a block diagram illustrating an apparatus for confirming a tv station icon according to an exemplary embodiment of the present invention;
fig. 7 is a block diagram of an analysis module 62 of an apparatus for confirming a tv station icon according to an exemplary embodiment of the present invention;
fig. 8 is a block diagram of a discard module 63 of an apparatus for confirming a tv station icon according to an exemplary embodiment of the present invention;
fig. 9 is a block diagram of a statistic module 64 of an apparatus for confirming a tv station icon according to an exemplary embodiment of the present invention.
Detailed Description
The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.
The television station icon is one of the special visual contents in the broadcast video, contains important semantic information of the television station such as the station name, the type, the copyright and the like, is an important identifier for distinguishing the broadcast television channels, and the identification of the station icon has important significance for program guide, content analysis, retrieval and the like of the broadcast television. Currently, in the identification of station icons, it is often encountered that station icons appear in complex background images, resulting in the appearance of extreme conditions, known as "image noise" in some cases. Especially in the case of a small number of samples, individual "image noise" can seriously affect the accuracy of identifying the tv station icon. At present, the number of television stations is very large, at least three hundred channels are counted, the similarity of some television station icons is higher, and the difficulty of accurately identifying the television station icons is increased.
Fig. 1 is a flowchart illustrating a method of confirming a station icon according to an exemplary embodiment, as shown in fig. 1, the method of confirming a station icon includes the following steps S11-S15:
in step S11, acquiring n frame images of a television station to be detected;
the method comprises the steps of extracting images of a television station to be detected, wherein the number of the extracted images is n frames, and the n frames of images can be continuous or discontinuous. All n frame images are images from the same television station.
In step S12, similar tv station icons ranked m before the similar probability of each frame image in the n frames of images of the tv station to be detected are analyzed, respectively, where the similar tv station icons include similar probabilities;
analyzing similar television station icons of an ith frame image of a television station to be detected, wherein the similar television station icons comprise similar probabilities, confirming that the television station icon m before the similar probability ranking of the ith frame image of the television station to be detected is the similar television station icon, traversing each frame image in the n frame images of the television station to be detected, and analyzing the similar television station icon m before the similar probability ranking of each frame image in the n frame images of the television station to be detected.
In step S13, the similarity probability of an anomaly is discarded;
firstly, counting the cumulative similarity probability sum of the jth similar television station icon; then, calculating the average similarity probability of the jth similar television station icon; then, calculating the variance of each similarity probability of the jth similar television station icon and the average similarity probability of the jth similar television station icon; then, judging whether the variance is larger than a preset variance threshold value; and finally, when the variance is smaller than a preset variance threshold value, confirming that the similarity probability corresponding to the variance is abnormal similarity probability, and abandoning the abnormal similarity probability.
And traversing all similar television station icons, and discarding the abnormal similar probabilities of all similar television station icons.
In step S14, after discarding the similarity probability of the anomaly, counting the cumulative sum of similarity probabilities for each similar tv station icon;
after discarding the abnormal similarity probability, the cumulative similarity probability sum of the jth similar TV station icon is counted again. By the scheme, all the icons of the similar television stations are traversed, and the cumulative similarity probability sums of all the similar television stations are respectively counted.
In step S15, the cumulative similarity probability and the highest similar station icon are determined to be the matching station icon.
And determining the accumulated similarity probability and the highest similar television station icon as a matched television station icon.
If the accumulated similarity probability and the highest similar television station icon are multiple (including the situation that the difference between the accumulated similarity probability and the highest similar television station icon is extremely small), the matched television station icon can be further determined according to the audio information corresponding to the n frames of images of the television station to be detected. And searching the audio information with the highest similarity to the audio information to be detected in a preset audio database. And utilizing the searched audio information, storing an information table of the corresponding relation between the audio information and the television station icon in preset audio data, and searching the matched television station icon according to the information table.
In one embodiment, as shown in FIG. 2, step S12 includes the following steps S21-S23:
in step S21, analyzing similar tv station icons of the ith frame image of the tv station to be detected, where the similar tv station icons include a similar probability;
analyzing any frame image in n frame images of the television station to be detected, and not assuming that the any frame image is the ith frame image. In a database of preset television station icons, g television station icons are stored in the database, and the g television station icons are standard television station icons. G similar television station icons and the similar probabilities corresponding to the g similar icons can be obtained by analyzing the similar probabilities of the ith frame image and the g television station icons in a preset database, and the similar probabilities are not named as Pi1、Pi2、……、Pig。
In step S22, confirming that the tv station icon m before the similarity probability ranking of the ith frame image of the tv station to be detected is a similar tv station icon;
by the similarity probability P of the ith frame image of the TV station to be detectedi1、Pi2、……、PigThe sequencing of (1) can sequence the top m similar TV station icons according to the size of the probability numerical value, and the similar probability set of the similar TV station icons which are not named the ith frame image of the TV station to be detected and are ranked in the top m is Pimaxm()。
In step S23, each frame image in the n frame images of the tv station to be detected is traversed, and the similar tv station icon m before the similar probability ranking of each frame image in the n frame images of the tv station to be detected is analyzed.
In the database of the preset television station icons, a similar probability set of each frame image and g television station icons in the preset database can be obtained by traversing each frame image in n frame images of the television station to be detected, and the similar probability set is not named asBy sorting the similarity probability of each frame image of the television station to be detected, the probability value can be largeThe similar television station icons in the top m of each frame image are sorted in a small way, and the similar probability set of the similar television station icons in the top m of the similar probability ranking of each frame image of the television station to be detected is not named as
In one embodiment, as shown in FIG. 3, step S13 includes the following steps S31-S37:
in step S31, counting the cumulative similarity probability sum of the jth similar tv station icon;
the similarity probability set of the similar television station icons m before the similarity probability ranking of each frame image of the television station to be detected isIn, do not assumeThe number of the similarity probability related to the jth similar TV station icon is t, and the t is countedThe sum of the similar probabilities related to the jth similar TV station icon can be obtained as the sum of the similar probabilities of the jth similar TV station icon, and the sum of the similar probabilities of the jth similar TV station icon is not named as Pj。
For example, the 14 th station icon in the database of preset station icons, does not assume29 of them are associated with the 14 th station icon, and the similar probabilities are superposed to obtain the similar probability and P of the 14 th station icon14。
In step S32, calculating an average similarity probability of the jth similar tv station icon;
p obtained as described abovejAnd t, calculating PjValue of division by tDo not name the value as the mean likelihood
In step S33, calculating a variance of each similarity probability of the jth similar station icon and an average similarity probability of the jth similar station icon;
is not supposed to be atThe number of the similarity probabilities related to the jth similar television station icon is t, and the similarity probability and the average similarity probability related to the jth similar television station icon are calculatedThe variance of (c).
In step S34, determining whether the variance is greater than a preset variance threshold;
the variance threshold is not assumed to be σ, and the above-mentioned value of the variance is compared with the variance threshold σ.
For example, 30 frames of images have a similarity probability with the 17 th television station icon, wherein the average similarity probability corresponding to the 30 frames of images is 92%, and the preset variance threshold is 30%, the variance between the similarity probability and the average similarity probability of each image in the 30 frames of images is calculated respectively, and whether the variance is greater than the preset variance threshold is judged.
In step S35, when the variance is greater than a preset variance threshold, confirming that a similarity probability corresponding to the variance is the abnormal similarity probability;
when the variance is greater than the value of σ, it is determined that the similarity probability corresponding to the variance is an abnormal similarity probability, that is, the image of the television station to be detected corresponding to the abnormal similarity probability is greatly different from most of the images of the television station to be detected, so that the similarity probability of the image detection is abnormal, and such an image is called as "image noise" in the image detection.
In the above example, there is a similarity probability of 25% corresponding to one frame of image, and the variance between the similarity probability of 25% and the average similarity probability of 92% is calculated to be much larger than the preset variance threshold of 30%, and the image with the similarity probability of 25% is called "image noise".
In step S36, discarding the similarity probability of the anomaly;
the above-mentioned similarity probability of an anomaly affects the subsequent calculation of the similarity probability, and the similarity probability of the anomaly is discarded.
For example, an image having a similarity probability of 25% in the above example is an abnormal similarity probability. That is, similar probabilities that differ significantly from the sample mean probabilities are discarded.
In step S37, all of the similar tv station icons are traversed, and the similarity probabilities of the anomalies for all of the similar tv station icons are discarded.
Make statistics ofThe number of similar television station icons and the corresponding similar probabilities of the similar television station icons. The method finds out the abnormal similar probability in each similar television station icon, and discards the abnormal similar probability in all similar television station icons.
In one embodiment, as shown in FIG. 4, step S14 includes the following steps S41-S42:
in step S41, after discarding the abnormal similar probabilities, counting again the cumulative similar probability sum of the jth similar tv station icon;
is abandonedAfter the abnormal similar probability of all the similar TV station icons, the similar probability set of the similar TV station icons m before the similar probability ranking of each frame image of the TV station to be detected is not called asIs not supposed to be atThe number of the similarity probability related to the jth similar TV station icon is r, and the r is countedThe sum of the similar probabilities related to the jth similar TV station icon can obtain the sum of the similar probabilities of the jth similar TV station icon, and the sum of the similar probabilities of the jth similar TV station icon is not named as Qj。
For example, the 31 st station icon in the database of preset station icons, is not assumed to beThere are 19 of these with similar probabilities associated with the 31 st station icon. Where there are 2 similar probabilities of anomalies. After discarding the similarity probabilities of these 2 anomalies,the similar probabilities of 17 television station icons and 31 st television station icon are superposed to obtain the similar probability and Q of the 31 st television station icon14。
In step S42, all the icons of the similar stations are traversed, and the cumulative sum of the similarity probabilities of all the similar stations is respectively counted.
After discarding all the similar probabilities of the anomalies, a result is formedIn thatThe method can obtain the types of the icons of the similar television stations and the similarity of each iconHow many similar probabilities a station icon corresponds to. And traversing all kinds of similar television station icons, and superposing all the similar probabilities corresponding to the same kind of similar television station icons to obtain the accumulated similar probability sum of all kinds of similar television stations.
For example,there are 4 kinds of similar tv station icons, and the first kind of similar tv station icons have 5 similar probabilities of 81%, 88%, 78%, 91%, and 84%, respectively; the existence of 4 similar probabilities for a second category of similar station icons, 67%, 73%, 69%, 68%, respectively; the existence of 6 similar probabilities for similar tv station icons of the third category, 95%, 94%, 93%, 97%, 90%, 88%, respectively; the existence of 6 similar probabilities for similar tv station icons of the fourth category is 98%, 96%, 95%, 96%, 94%, 98%, respectively. Then the sum of the similarity probabilities for similar tv station icons of each category can be calculated. The sum of similarity probabilities for similar tv station icons of the fourth category may be calculated to be 577%, which is also the highest cumulative sum of similarity probabilities. That is, a similar station icon of the fourth category is determined to be a matching station icon.
In one embodiment, as shown in FIG. 5, step S22 includes the following steps S51-S55:
in step S51, determining whether all the similarity probabilities m before the similarity probability ranking of the ith frame image of the tv station to be detected are greater than or equal to a preset limited probability threshold;
and if the preset limited probability threshold value is not named as gamma, judging the numerical value of the similarity probability of the ith frame image of the television station to be detected before the ranking m and the numerical value of the gamma.
In step S52, when the determination is yes, confirming that the tv station icon m before the similarity probability ranking of the ith frame image of the tv station to be detected is a similar tv station icon;
and when all the similarity probabilities in the m-th similarity probability ranking of the ith frame image of the television station to be detected are more than or equal to gamma, confirming that the television station icon in the m-th similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon.
In step S53, when the determination is negative, comparing to find out the similar tv station icon at k before the similar probability of the ith frame image of the tv station to be detected is greater than or equal to the preset limited probability threshold, where k is smaller than m;
when the partial similarity probability in the m-th similarity probability ranking of the similarity probability of the ith frame image of the television station to be detected is greater than gamma, comparing the partial similarity probability greater than gamma, and for convenience of expression, assuming that the number of the partial similarity probability is k.
In step S54, determining that the tv station icon k before the similarity probability ranking of the ith frame image of the tv station to be detected is a similar tv station icon;
the effect of the predetermined threshold value γ is to exclude too low a similarity probability in order to reduce adverse effects on subsequent steps. In case the similarity probability is below a pre-set defined probability threshold γ, the degree of similarity is considered too low. In this case, the television station icon with the similarity probability of the ith frame image of the television station to be detected being k before the ranking is determined as the similar television station icon.
For example, if the preset defined probability threshold γ is not assumed to be 50%, if the similarity probability of the 5 th frame image of the tv station to be detected ranks 5 top, and the similarity probabilities are 91%, 88%, 78%, 75% and 34%, then 34% may be excluded from comparison with the defined probability threshold 50%, and a similarity probability lower than 50% is equivalent to a similarity degree that is too low.
In step S55, when k is equal to 0, it is determined that the ith frame image of the tv station to be detected does not have a corresponding similar tv station icon.
In one embodiment, when the similarity probability of the ith frame image of all the television stations to be detected is that the similarity probability of m before the ranking is smaller than the preset limited probability threshold value gamma, determining that the ith frame image of the television station to be detected does not have a corresponding similar television station icon.
In one embodiment, fig. 6 is a block diagram illustrating an apparatus for confirming a television station icon in accordance with an exemplary embodiment. As shown in fig. 6, the apparatus includes an acquisition module 61, an analysis module 62, a rejection module 63, a statistics module 64, and a determination module 65.
The acquiring module 61 is used for acquiring n frame images of a television station to be detected;
the analysis module 62 is configured to analyze similar tv station icons m before the ranking of the similar probability of each frame image in the n frames of images of the tv station to be detected, where the similar tv station icons include similar probabilities;
the discarding module 63 is configured to discard the similarity probability of an anomaly;
the statistical module 64 is configured to, after discarding the abnormal similarity probabilities, perform statistics on cumulative similarity probability sums of the similar tv station icons;
the determining module 65 is configured to determine the cumulative similarity probability and the highest similar station icon as the matching station icon.
As shown in fig. 7, the analysis module 62 includes an analysis sub-module 71, a first confirmation sub-module 72, and a first traversal sub-module 73.
The analysis submodule 71 is configured to analyze similar tv station icons of an ith frame image of the tv station to be detected, where the similar tv station icons include a similar probability;
the first confirming submodule 72 is configured to confirm that the tv station icon m before the similarity probability ranking of the ith frame image of the tv station to be detected is a similar tv station icon;
the first determining submodule 72 is further configured to determine whether all the similarity probabilities m before the similarity probability ranking of the ith frame image of the television station to be detected are greater than or equal to a preset limited probability threshold; when the judgment is yes, confirming that the television station icon m before the similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon; if not, comparing the similar television station icons with the similarity probability of the ith frame image of the television station to be detected being greater than or equal to the similarity probability ranking k before the preset limited probability threshold, wherein k is smaller than m; calculating the similar television station icon of the k before the similarity probability ranking of the ith frame image of the television station to be detected; and when the k is equal to 0, determining that the ith frame image of the television station to be detected does not have a corresponding similar television station icon.
The first traversal submodule 73 is configured to traverse each frame image in the n frames of images of the television station to be detected, and analyze a similar television station icon m before the similarity probability ranking of each frame image in the n frames of images of the television station to be detected.
As shown in FIG. 8, the discard module 63 includes a first statistics submodule 81, a first calculation submodule 82, a second calculation submodule 83, a decision submodule 84, a second validation submodule 85, a discard submodule 86, and a second traversal submodule 87.
The first statistic submodule 81 is configured to count a cumulative similarity probability sum of the jth similar tv station icon;
the first calculating submodule 82 is configured to calculate an average similarity probability of the jth similar tv station icon;
the second calculating submodule 83 is configured to calculate a variance between each similarity probability of the jth similar tv station icon and an average similarity probability of the jth similar tv station icon;
the judgment submodule 84 is configured to judge whether the variance is greater than a preset variance threshold;
the second determining submodule 85 is configured to determine that the similarity probability corresponding to the variance is the abnormal similarity probability when the variance is greater than a preset variance threshold;
the discard submodule 86, configured to discard the similarity probability of the anomaly;
the second traversal submodule 87 is configured to traverse all of the similar tv station icons and discard the abnormal similar probabilities of all of the similar tv station icons.
As shown in fig. 9, the statistics module 64 includes a second statistics submodule 91 and a third traversal submodule 92.
The second statistic submodule 91 is configured to count the cumulative sum of the similarity probabilities of the jth similar tv station icon again after discarding the abnormal similarity probabilities;
the third traversal submodule 92 is configured to traverse all the icons of the similar tv stations, and count the cumulative sum of the similarity probabilities of all the similar tv stations respectively.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (6)
1. A method for confirming a television station icon, comprising:
acquiring n frame images of a television station to be detected;
respectively analyzing similar television station icons m before the similarity probability of each frame image in n frames of images of the television station to be detected, wherein the similar television station icons comprise similar probabilities, any frame image is an ith frame image, g television station icons are stored in a preset database of the television station icons, the g television station icons are standard television station icons, and g television station icons and the similar probabilities corresponding to the g similar television station icons are obtained by analyzing the similarity probabilities of the ith frame image and the g television station icons in the preset database;
discarding the similarity probability of an anomaly;
after discarding the abnormal similar probability, counting the cumulative similar probability sum of each similar TV station icon;
determining the accumulated similarity probability and the highest similar TV station icon as a matched TV station icon;
the analyzing of the similar television station icons m before the ranking of the similar probability of each frame image in the n frames of images of the television station to be detected respectively comprises the following steps:
analyzing similar TV station icons of the ith frame image of the TV station to be detected, wherein the similar TV station icons comprise similar probability;
confirming that the television station icon m before the similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon;
traversing each frame image in the n frames of images of the television station to be detected, and analyzing similar television station icons m before the similar probability ranking of each frame image in the n frames of images of the television station to be detected;
said similarity probabilities of said rejection anomalies comprising:
counting the cumulative similarity probability sum of the jth similar TV station icon, wherein the similarity probability set of the similar TV station icons m before the similarity probability ranking of each frame image of the TV station to be detected isSuppose inThe number of the similarity probability related to the jth similar TV station icon is r, and the r is countedObtaining the sum of the similar probabilities related to the jth similar TV station icon;
calculating the average similarity probability of the jth similar television station icon;
calculating the variance of each similarity probability of the jth similar television station icon and the average similarity probability of the jth similar television station icon;
judging whether the variance is larger than a preset variance threshold value or not;
when the variance is larger than a preset variance threshold value, confirming that the similarity probability corresponding to the variance is the abnormal similarity probability;
discarding the likelihood probability of the anomaly;
traversing all of the similar TV station icons, and discarding the abnormal similar probabilities of all of the similar TV station icons.
2. The method of claim 1, wherein said counting a cumulative sum of similarity probabilities for each similar station icon after discarding said similarity probabilities for said anomalies comprises:
after the abnormal similar probability is abandoned, the cumulative similar probability sum of the jth similar TV station icon is counted again;
and traversing all the icons of the similar television stations, and respectively counting the cumulative similarity probability sums of all the similar television stations.
3. The method of claim 1, wherein said confirming that the tv station icon m before the similarity probability of the ith frame image of the tv station to be detected is a similar tv station icon comprises:
judging whether the similarity probabilities m before the similarity probability ranking of the ith frame image of the television station to be detected are all larger than or equal to a preset limited probability threshold value;
when the judgment is yes, confirming that the television station icon m before the similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon;
if not, comparing the similar television station icons with the similarity probability of the ith frame image of the television station to be detected being greater than or equal to the similarity probability ranking k before the preset limited probability threshold, wherein k is smaller than m;
calculating the similar television station icon of the k before the similarity probability ranking of the ith frame image of the television station to be detected;
and when the k is equal to 0, determining that the ith frame image of the television station to be detected does not have a corresponding similar television station icon.
4. An apparatus for confirming a television station icon, comprising:
the acquisition module is used for acquiring n frame images of a television station to be detected;
the analysis module is used for respectively analyzing similar television station icons m before the similar probability ranking of each frame image in the n frames of images of the television station to be detected, wherein the similar television station icons comprise similar probabilities, any frame image is an ith frame image, g television station icons are stored in a preset database of the television station icons, the g television station icons are standard television station icons, and g television station icons and the similar probabilities corresponding to the g television station icons are obtained by analyzing the similar probabilities of the ith frame image and the g television station icons in the preset database;
a discarding module for discarding the similarity probability of an anomaly;
the statistic module is used for counting the cumulative similarity probability sum of each similar TV station icon after discarding the abnormal similarity probability;
a determining module for determining the accumulated similarity probability and the highest similar TV station icon as a matching TV station icon;
the analysis module comprises:
the analysis submodule is used for analyzing similar television station icons of the ith frame image of the television station to be detected, and the similar television station icons comprise similar probabilities;
the first confirming submodule is used for confirming that the television station icon m before the similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon;
the first traversal submodule is used for traversing each frame image in the n frame images of the television station to be detected and analyzing similar television station icons m before the similar probability ranking of each frame image in the n frame images of the television station to be detected;
the discard module comprises:
a first statistic submodule for counting the cumulative similarity probability sum of the jth similar TV station icon, wherein the similarity probability set of the similar TV station icons m before the similarity probability ranking of each frame image of the TV station to be detected isSuppose inThe number of the similarity probability related to the jth similar TV station icon is r, and the r is countedObtaining the sum of the similar probabilities related to the jth similar TV station icon;
the first calculation submodule is used for calculating the average similarity probability of the jth similar television station icon;
the second calculation submodule is used for calculating the variance between each similarity probability of the jth similar television station icon and the average similarity probability of the jth similar television station icon;
the judgment submodule is used for judging whether the variance is larger than a preset variance threshold value or not;
the second confirming submodule is used for confirming that the similarity probability corresponding to the variance is abnormal when the variance is larger than a preset variance threshold;
a discard submodule for discarding the likelihood probability of the anomaly;
and the second traversal submodule is used for traversing all the similar television station icons and abandoning the abnormal similar probability of all the similar television station icons.
5. The apparatus of claim 4, wherein the statistics module comprises:
the second statistic submodule is used for counting the cumulative similarity probability sum of the jth similar TV station icon again after the abnormal similarity probability is discarded;
and the third traversal submodule is used for traversing all the icons of the similar television stations and respectively counting the cumulative similarity probability sum of all the similar television stations.
6. The apparatus of claim 4,
the first confirming submodule is also used for judging whether the similarity probabilities m before the similarity probability ranking of the ith frame image of the television station to be detected are all larger than or equal to a preset limited probability threshold value; when the judgment is yes, confirming that the television station icon m before the similarity probability ranking of the ith frame image of the television station to be detected is a similar television station icon; if not, comparing the similar television station icons with the similarity probability of the ith frame image of the television station to be detected being greater than or equal to the similarity probability ranking k before the preset limited probability threshold, wherein k is smaller than m; calculating the similar television station icon of the k before the similarity probability ranking of the ith frame image of the television station to be detected; and when the k is equal to 0, determining that the ith frame image of the television station to be detected does not have a corresponding similar television station icon.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610124860.5A CN107153809B (en) | 2016-03-04 | 2016-03-04 | Method and device for confirming television station icon |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610124860.5A CN107153809B (en) | 2016-03-04 | 2016-03-04 | Method and device for confirming television station icon |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107153809A CN107153809A (en) | 2017-09-12 |
CN107153809B true CN107153809B (en) | 2020-10-09 |
Family
ID=59792313
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610124860.5A Expired - Fee Related CN107153809B (en) | 2016-03-04 | 2016-03-04 | Method and device for confirming television station icon |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107153809B (en) |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2004086751A2 (en) * | 2003-03-27 | 2004-10-07 | Sergei Startchik | Method for estimating logo visibility and exposure in video |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN100399813C (en) * | 2005-03-31 | 2008-07-02 | 联想(北京)有限公司 | Method for realizing channel automatic correspondence based on electronic program guide |
JP5364528B2 (en) * | 2009-10-05 | 2013-12-11 | 株式会社日立ハイテクノロジーズ | Pattern matching method, pattern matching program, electronic computer, electronic device inspection device |
-
2016
- 2016-03-04 CN CN201610124860.5A patent/CN107153809B/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2004086751A2 (en) * | 2003-03-27 | 2004-10-07 | Sergei Startchik | Method for estimating logo visibility and exposure in video |
Non-Patent Citations (2)
Title |
---|
A Well-Behaved TV Logo Recognition Method Using Heuristic Optimality Checked Basis Pursuit Denoising;Jie Xu等;《ICIG 2015: Image and Graphics》;20150804;363-367 * |
基于HSV颜色空间加权Hu不变矩的台标识别;周献中等;《南京理工大学学报》;20050630;第29卷(第3期);529-540 * |
Also Published As
Publication number | Publication date |
---|---|
CN107153809A (en) | 2017-09-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8358837B2 (en) | Apparatus and methods for detecting adult videos | |
CN102254006B (en) | Method for retrieving Internet video based on contents | |
CN109815991A (en) | Training method, device, electronic equipment and the storage medium of machine learning model | |
CN111090776A (en) | Video auditing method, device, auditing server and storage medium | |
EP2471025B1 (en) | A method and system for preprocessing the region of video containing text | |
CN111310826B (en) | Method and device for detecting labeling abnormality of sample set and electronic equipment | |
CN104881675A (en) | Video scene identification method and apparatus | |
CN107609149B (en) | Video positioning method and device | |
US11341747B2 (en) | Generation of video hash | |
CN111932394A (en) | Bad root cause path analysis method and system based on association rule mining | |
CN111210402A (en) | Face image quality scoring method and device, computer equipment and storage medium | |
CN113435443B (en) | Method for automatically identifying landmark from video | |
CN108052931A (en) | A kind of license plate recognition result fusion method and device | |
Hanjalic et al. | Template-based detection of anchorperson shots in news programs | |
CN111368867A (en) | Archive classification method and system and computer readable storage medium | |
CN111126112B (en) | Candidate region determination method and device | |
CN111914649A (en) | Face recognition method and device, electronic equipment and storage medium | |
US9699501B2 (en) | Information processing device and method, and program | |
CN111541939A (en) | Video splitting method and device, electronic equipment and storage medium | |
US11682211B2 (en) | Computer-implemented method | |
CN107153809B (en) | Method and device for confirming television station icon | |
CN112818984B (en) | Title generation method, device, electronic equipment and storage medium | |
CN110633466B (en) | Short message crime identification method and system based on semantic analysis and readable storage medium | |
US9020268B2 (en) | Method and apparatus for annotating multimedia data in a computer-aided manner | |
US11386340B2 (en) | Method and apparatus for performing block retrieval on block to be processed of urine sediment image |
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 | ||
PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Method and device for confirming TV station Icon Effective date of registration: 20210104 Granted publication date: 20201009 Pledgee: Inner Mongolia Huipu Energy Co.,Ltd. Pledgor: WUXI TVMINING MEDIA SCIENCE & TECHNOLOGY Co.,Ltd. Registration number: Y2020990001517 |
|
PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20201009 Termination date: 20210304 |
|
CF01 | Termination of patent right due to non-payment of annual fee |