CN111967312A - Method and system for identifying important persons in picture - Google Patents
Method and system for identifying important persons in picture Download PDFInfo
- Publication number
- CN111967312A CN111967312A CN202010641879.3A CN202010641879A CN111967312A CN 111967312 A CN111967312 A CN 111967312A CN 202010641879 A CN202010641879 A CN 202010641879A CN 111967312 A CN111967312 A CN 111967312A
- Authority
- CN
- China
- Prior art keywords
- person
- picture
- characteristic value
- character
- value
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 55
- 238000005516 engineering process Methods 0.000 claims abstract description 10
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 230000009286 beneficial effect Effects 0.000 abstract description 7
- 238000011160 research Methods 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 12
- 230000008569 process Effects 0.000 description 9
- 230000000694 effects Effects 0.000 description 8
- 230000009471 action Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/70—Multimodal biometrics, e.g. combining information from different biometric modalities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method and a system for identifying important persons in pictures, and relates to the technical field of picture identification. The method comprises the steps of obtaining basic information data of an identified object in a picture through an ImageAI technology; then, acquiring a character face characteristic value, a character size characteristic value and a character position characteristic value according to the identification object basic information data of which the types belong to the characters; and finally, sorting the importance scores, and identifying the important persons in the picture by comparing the importance scores with a preset threshold value. The method can not only sort the characters according to the importance, but also accurately identify the important characters in the picture according to the threshold value, and is beneficial to the research of related tasks such as image description and the like.
Description
Technical Field
The invention relates to the technical field of picture identification, in particular to a method and a system for identifying important persons in pictures.
Background
Since the picture can realize the information transmission without language barrier, it becomes one of important media for transmitting and sharing information in the present society. As shown in fig. 1, no matter which language people use the above illustration, it indicates that the picture is that two basketball players are playing the game, and if more information is known to the two people, it can further indicate that "the 11 th yaoming of houston rocket team and the 24 th science of los angeles lake team are playing the basketball game", so that the picture can transmit information without language obstacle and can be used as an important means for acquiring information. Identifying important persons in a picture means that, given a picture containing many figures, the important persons in the picture are identified. The important person is a person which is in a prominent position in the picture, plays a more important role than other persons in the picture, is closely related to the content of the picture, and the name of the person usually appears in the text corresponding to the picture, and the important person in the picture is closely related to the subject of the picture.
In the existing method, the important persons in the pictures are mainly identified by using an Image AI technology, and the Image AI is a powerful open source Python (Python is a computer programming language) project library which contains machine learning algorithms related to various computer vision tasks, can support tasks in various vision fields such as picture object identification, picture category prediction, video object detection and tracking, and is widely used by research groups and individuals all over the world.
However, the inventors of the present application found that the probability value given by the ImageAI technique is a probability that the target object is a person, and the value does not reflect the degree of importance of the person. Namely, the method cannot accurately identify important characters in the picture.
Disclosure of Invention
Technical problem to be solved
Aiming at the defects of the prior art, the invention provides a method and a system for identifying important persons in a picture, which solve the technical problem that the important persons in the picture cannot be accurately identified in the prior art.
(II) technical scheme
In order to achieve the purpose, the invention is realized by the following technical scheme:
the invention provides a method for identifying important persons in pictures, which comprises the following steps:
acquiring basic information data of a picture recognition object based on an ImageAI technology;
acquiring a character face characteristic value, a character size characteristic value and a character position characteristic value based on the basic information data;
acquiring an importance score of the person based on the person face characteristic value, the person size characteristic value and the person position characteristic value;
and sorting the importance scores, and identifying important persons in the picture by comparing the importance scores with a preset threshold value.
Preferably, the basic information data includes: category, probability value, and boundary position coordinates.
Preferably, the obtaining of the human face feature value based on the basic information data includes:
wherein, FaceiRepresenting the face feature value of the ith person.
Preferably, the obtaining of the size characteristic value of the person based on the basic information data includes:
wherein, SizeiA size characteristic value representing the ith person; person _ areaiThe character frame area of the i-th character is shown.
Preferably, the obtaining of the person position feature value based on the basic information data includes:
the character position characteristic value comprises a character horizontal center characteristic value and a character vertical center characteristic value;
the calculation formula of the character horizontal center characteristic value is as follows:
wherein, Position _ HiHorizontal center feature value, 0, representing the ith individual<Position_HiLess than or equal to 1; picture _ H represents the distance of the Picture horizontal center from the Picture on-Picture boundary; personiH represents the horizontal center distance of the character frame of the ith personDistance of the border on the picture;
the figure vertical center characteristic value calculation formula is as follows:
wherein: position _ ViVertical center feature value, 0, for the ith individual<Position_ViLess than or equal to 1; picture _ V denotes the distance of the vertical center of a Picture from the left boundary of the Picture, PersoniAnd _Vrepresents the distance of the vertical center of the character frame of the ith person from the left border of the picture.
Preferably, the obtaining an importance score of the person based on the face feature value, the size feature value and the position feature value of the person, and identifying an important person in the picture according to the importance score of the person includes:
the formula for calculating the importance score of a person is as follows:
Importance_Score＝α1*Face+α2*Size+α3*Position_H+α4*Position_V
wherein, Importance _ Score represents the Importance Score of the person in the picture, 0<Importance_Score≤1；α1、α2、αα、α4Respectively, the weights of the four features.
Preferably, the preset threshold is 0.7.
The invention also provides a system for identifying important persons in pictures, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the steps of the method when executing the computer program.
(III) advantageous effects
The invention provides a method and a system for identifying important persons in pictures. Compared with the prior art, the method has the following beneficial effects:
the method comprises the steps of obtaining basic information data of an identified object in a picture through an ImageAI technology; then, acquiring a character face characteristic value, a character size characteristic value and a character position characteristic value according to the identification object basic information data of which the types belong to the characters; and finally, sorting the importance scores, and identifying the important persons in the picture by comparing the importance scores with a preset threshold value. The method can not only sort the characters according to the importance, but also accurately identify the important characters in the picture according to the threshold value, and is beneficial to the research of related tasks such as image description and the like.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a diagram of the effect of ImageAI on the recognition of a picture;
FIG. 2 is a diagram of the effect of ImageAI on the recognition of a picture;
FIG. 3 is a diagram of the effect of ImageAI on the recognition of a picture;
FIG. 4 is a diagram of the effect of ImageAI on the recognition of a picture;
FIG. 5 is a block diagram of a method for identifying important persons in a picture according to an embodiment of the present invention;
FIG. 6 is a diagram of the effect of ImageAI on the recognition of a picture;
FIG. 7 is a schematic diagram illustrating a value of a feature value of a human face according to an embodiment of the present invention;
FIG. 8 is a schematic diagram illustrating a value of a feature value of a human face according to an embodiment of the present invention;
FIG. 9 is a schematic diagram illustrating a figure size characteristic value in an embodiment of the present invention;
FIGS. 10A and 10B are schematic diagrams illustrating division of a horizontal center and a vertical center in an embodiment of the present invention;
fig. 11 is a schematic view of a value of a horizontal center feature value in feature values of a position of a person according to an embodiment of the present invention;
fig. 12 is a schematic diagram of a value of a vertical center feature value in feature values of a position of a person according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described, and it is obvious that the described embodiments are a part of the embodiments of the present invention, but not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the application provides a method for identifying important persons in pictures, so that the technical problem that the important persons in the pictures cannot be accurately identified in the prior art is solved, and the important persons in the pictures can be accurately identified.
In order to solve the technical problems, the general idea of the embodiment of the application is as follows:
in the existing method, the important persons in the picture are mainly identified by the ImageAI technology, but the ImageAI technology has the following defects when identifying the important persons in the picture:
(1) can not directly judge whether the identification result is an important figure
The probability value given by the ImageAI technique is the probability that the target object is a person, and the value does not reflect the importance of the person. As shown in fig. 2 and 3: the three persons on the prize table of fig. 2 are important persons, the probability value of the lowest person among the three persons is 88.064%, and there are two persons in fig. 3, the probability value of the lower person among the two persons is 90.593%, and they apparently do not belong to the important person, so that the magnitude of the probability value of the persons is irrelevant to whether it is the important person or not, as can be seen from the comparison of the probability values.
(2) The person recognition result of ImageAI is not one hundred percent correct. Referring to fig. 4: for example, there are four persons in fig. 4, but the ImageAI recognition result is five persons. Therefore, in the prior art, the identification effect cannot be guaranteed when identifying important persons, and therefore, a method capable of accurately identifying important persons in pictures is urgently needed.
In order to better understand the technical solution, the technical solution will be described in detail with reference to the drawings and the specific embodiments.
An embodiment of the present invention provides a method for identifying an important person in a picture, where the method is executed by a computer, and as shown in fig. 5, the method includes steps S1 to S4:
s1, acquiring basic information data of the picture recognition object based on the ImageAI technology;
s2, acquiring a character face characteristic value, a character size characteristic value and a character position characteristic value based on the basic information data;
s3, acquiring importance scores of the persons based on the face feature values, the size feature values and the position feature values of the persons;
and S4, sorting the importance scores, and identifying important persons in the picture by comparing the importance scores with a preset threshold value.
The embodiment of the invention not only can sort the people according to the importance, but also can accurately identify the important people in the picture according to the threshold value, and is beneficial to the research of related tasks such as image description and the like.
In one embodiment, the basic information data of the picture recognition object is acquired based on the ImageAI technology. The specific implementation process is as follows:
the basic information data includes its category, probability value, and boundary position coordinates.
ImageAI provides three trained models for picture object recognition (RetinaNet, YOLOv3, TinyYOLOv3), all of which can be downloaded directly in the object recognition module (Objectdetection) of its homepage. On the premise that ImageAI is installed in a computer, the process of identifying the figure is specifically realized as follows:
(1) creating a python file, downloading a trained object recognition model file, and placing the model file and the trained object recognition model file in the same folder;
(2) importing an object detection module and an os module in a python file, and acquiring a working path execution _ path of the current file;
(3) defining an ObjectDetection type detector, setting a called model type and specifying a model path, and finally loading a model to prepare object identification;
(4) and saving the recognition result graph, and outputting basic information data of each recognition object, including the category, the probability value and the boundary position coordinates of the recognition object.
Fig. 6 is a diagram showing the effect of image ai on identifying a person in a picture, where a box is a recognition boundary of a target object, the object is a "person" class, a right value is a probability that the object is the "person" class, and an upper coordinate value is distances from left, upper, right, and lower sides of the box to left, upper, left, and upper boundaries of the picture, and these values are all in units of pixel values.
In one embodiment, the person face feature value, the person size feature value, and the person position feature value are obtained based on the basic information data S2. The specific implementation process is as follows:
s201, acquiring a character face characteristic value based on basic information data, specifically comprising:
the important person in the picture is the core of the picture, and the face of the important person is often in the front or the side in the picture, so it is considered that the person appearing in the front or the side in the picture is very likely to be the important person, so the face feature value of the person in the picture is set as shown in formula (1):
wherein, FaceiRepresenting the face feature value of the ith person.
As shown in fig. 7, in the first case where both characters are sides, their face feature value is set to 1.
As shown in fig. 8, if there are both a front person and a side person in the figure, and the second case is, the face feature value of the front person is set to 1, and the face feature value of the side person is set to 0.5.
S202, acquiring a person Size characteristic value Size based on the basic information data. The method specifically comprises the following steps:
when there are a plurality of persons in one picture, the larger the size of the person frame (identified bounding box), the greater the probability that the person is an important person, and therefore the size characteristic value of the person frame is calculated as shown in formula (2):
wherein, SizeiA size characteristic value representing the ith character; person _ areaiThe area of the character frame representing the ith character can be calculated through the coordinates of the boundary position in the first step, the length and the width of the character frame can be obtained through the difference of the distances of all coordinates, and the length and the width are multiplied to obtain the area; and returning the maximum value of the area of the character frame in all the character recognition results by the max function of the denominator, specifically, obtaining the area of each character frame, and sequentially comparing until a final result is obtained.
As shown in fig. 9, the size characteristic value of the character 1 is 1.000, which is greater than 0.468 and 0.414 of the characters 2 and 3 around the character, and the probability that all the characters 1 are important characters is greater than that of the characters 2 and 3, which also corresponds to the actual situation of the figure.
S203, acquiring and calculating the characteristic value (divided into a horizontal center characteristic value Position _ H and a vertical center characteristic value Position _ V) of the Position of the person based on the basic information data. The method specifically comprises the following steps:
generally, the more important the person is, the closer the center position of the picture is.
As shown in fig. 10A and 10B, the horizontal straight line in the middle of the picture and the character frame represents the horizontal center of the entire picture and the character frame, and the vertical straight line in the middle of the picture and the character frame represents the vertical center of the entire picture and the character frame. The horizontal center feature of the person represents the degree of the horizontal center of the person frame approaching the horizontal center of the picture, the vertical center feature of the person represents the degree of the vertical center of the person frame approaching the vertical center of the picture, and the importance degree of the person in the picture can be judged according to the two features.
Position_HiRepresents the ithThe specific calculation formula of the horizontal center characteristic value of the human body is as follows:
wherein Picture _ H represents the distance of the horizontal center of the Picture from the boundary on the Picture, PersoniH represents the distance from the horizontal center of the character frame of the ith person to the boundary on the picture, Position HiThe value range of (a) is greater than 0 and less than or equal to 1.
The horizontal center feature values of the ten people in salutation as in the middle of fig. 11 are 0.9755, 0.9842, 0.9720, 0.9805, 0.9854, 0.9642, 0.9624, 0.9635, 0.9706, 0.9721 from left to right, respectively, and the probability that they are important people is much higher than the other people in the picture.
Position_ViThe vertical center characteristic value of the ith person is represented, and the specific calculation formula is as follows:
wherein Picture _ V represents the distance of the vertical center of the Picture from the left boundary of the Picture, PersoniV represents the distance from the vertical center of the character frame of the ith person to the left boundary of the picture, Position ViThe value range of (a) is greater than 0 and less than or equal to 1.
Taking fig. 12 as an example, the feature values of the vertical centers of the detected characters from left to right are 0.716, 0.99, and 0.766, respectively, and it can be seen from the results that the feature value of the middle character is higher than those of the left and right people, and thus the probability that the middle character is an important character is higher than those of the other people.
In one embodiment, the importance score of the person is obtained based on the face feature value of the person, the size feature value of the person, and the position feature value of the person S3. The specific implementation process is as follows:
the method for calculating the importance of the person in the picture based on the Face feature value Face, the Size feature value Size, the horizontal center feature value Position _ H and the vertical center feature value Position _ V of the person is shown in formula (5):
Importance_Score＝α1*Face+α2*Size+α3*Position_H+α4*Position_V(5)
the Importance _ Score value represents the Importance Score of the character in the picture, the value range of the Importance Score is larger than 0 and smaller than or equal to 1, and the larger the value of the Importance Score is, the larger the Importance degree of the corresponding character is; alpha is alpha1、α2、α3、α4(the sum of the four weights is 1) is the weight of each of the four features. In the present example, a comparison experiment was performed on a set of pictures using four sets of weight values of 0.25, 0.4, 0.2, 0.4, 0.2, and 0.2, 0.3, and 0.3, respectively, and as a result, as shown in table 1, it was found that the ratio of important persons (the number of important persons having a score of 0.7 or more/the number of all important persons, i.e., the accuracy) was highest at 94.53% with a score of 0.7 or more at the second set of weight values.
TABLE 1 important figure proportion score higher than 0.7 at different weight values
In one embodiment, the importance scores are sorted at S4, and the important persons in the picture are identified by comparing the importance scores with a preset threshold. The specific implementation process is as follows:
through experiments and observation of results, when the value of the Importance _ Score of a person is greater than or equal to 0.7, the person can be basically judged as an important person, so that all important persons in a picture are obtained through the threshold of 0.7, and the Importance degrees are sorted according to the size.
For the example of an incorrect ImageAI identification, as shown in fig. 4:
the importance scores of the individual character recognition results in fig. 4 calculated by the above method are shown in table 2:
TABLE 2
| Probability value of person | Importance score | Whether or not there is an important character |
| 92.405％ | 0.937 | Is that |
| 94.843％ | 0.850 | Is that |
| 92.261％ | 0.837 | Is that |
| 93.037％ | 0.810 | Is that |
| 59.969％ | 0.555 | Whether or not |
Through analysis, the recognition result with the figure probability value of 59.969% is judged to be wrong, and the result is analyzed from the picture to recognize the overlapped area of the two figures as the figures, so that the method can effectively detect the mistakes and reduce the influence of the wrong recognition result on the object recognition task.
In the embodiment of the invention, for pictures in a specific field, such as sports pictures, a face database of important sports persons can be established, and then the persons can be identified in a given picture by using a face identification tool.
It should be noted that the embodiment of the present invention can also be applied to research on name recognition of a language with scarce resources. The process mainly comprises the following steps: the method comprises the steps of processing images and corresponding text data in a scarce resource source language, identifying important characters in the images, obtaining names of people in a general language through face identification, translating the names to obtain names of people in a target language, and matching the names of people in an original text to obtain proper names of people.
The embodiment of the invention also provides a system for identifying important persons in the picture, which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the steps of the method are realized when the processor executes the computer program.
It can be understood that the system for identifying the important person in the picture provided by the embodiment of the present invention corresponds to the method for identifying the important person in the picture, and the explanation, examples, and beneficial effects of the relevant contents thereof may refer to the corresponding contents in the method for identifying the important person in the picture, which are not described herein again.
In summary, compared with the prior art, the method has the following beneficial effects:
1. the embodiment of the invention not only can sort the people according to the importance, but also can accurately identify the important people in the picture according to the threshold value, and is beneficial to the research of related tasks such as image description and the like.
2. The embodiment of the invention designs four indexes and calculation modes for judging the importance of the character in the picture, wherein the indexes are a character face characteristic value, a character size characteristic value, a horizontal center characteristic value and a vertical center characteristic value of the character, the value ranges of the four characteristic values are all between 0 and 1, four corresponding weights are set for the four indexes, 0.7 is used as a threshold value to divide the important character and the non-important character, when the character importance score is more than or equal to 0.7, the character is judged to be the important character, otherwise, the character is not judged. Compared with the existing ImageAI technology, on one hand, the embodiment of the invention can solve the problem that the ImageAI can not identify important persons, and on the other hand, the person identification result of the ImageAI can be analyzed by identifying the information of the important persons, thereby solving the problem that the person identification is wrong.
It should be noted that, through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
In this document, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (8)
1. A method for identifying important persons in pictures is characterized by comprising the following steps:
acquiring basic information data of a picture recognition object based on an ImageAI technology;
acquiring a character face characteristic value, a character size characteristic value and a character position characteristic value based on the basic information data;
acquiring an importance score of the person based on the person face characteristic value, the person size characteristic value and the person position characteristic value;
and sorting the importance scores, and identifying important persons in the picture by comparing the importance scores with a preset threshold value.
2. The method for identifying important persons in pictures according to claim 1, wherein the basic information data comprises: category, probability value, and boundary position coordinates.
3. The method for identifying an important person in a picture according to claim 2, wherein said obtaining a face feature value of the person based on the basic information data comprises:
wherein, FaceiRepresenting the face feature value of the ith person.
4. The method for identifying important persons in pictures as claimed in claim 2, wherein said obtaining the size characteristic value of the persons based on the basic information data comprises:
wherein, SizeiA size characteristic value representing the ith person; person _ areaiThe character frame area of the i-th character is shown.
5. The method for identifying important persons in pictures as claimed in claim 2, wherein said obtaining the position characteristic value of the person based on the basic information data comprises:
the character position characteristic value comprises a character horizontal center characteristic value and a character vertical center characteristic value;
the calculation formula of the character horizontal center characteristic value is as follows:
wherein, Position _ HiHorizontal center feature value, 0, representing the ith individual<Position_HiLess than or equal to 1; picture _ H represents the distance of the Picture horizontal center from the Picture on-Picture boundary; personiH represents the distance from the horizontal center of the character frame of the ith person to the upper boundary of the picture;
the figure vertical center characteristic value calculation formula is as follows:
wherein: position _ ViVertical center feature value, 0, for the ith individual<Position_ViLess than or equal to 1; picture _ V denotes the distance of the vertical center of a Picture from the left boundary of the Picture, PersoniAnd _Vrepresents the distance of the vertical center of the character frame of the ith person from the left border of the picture.
6. The method of claim 5, wherein the obtaining of the importance score of the person based on the face feature value, the size feature value and the position feature value of the person, and the identifying of the important person in the picture according to the importance score of the person comprises:
the formula for calculating the importance score of a person is as follows:
Importance_Score＝α1*Face+α2*Size+α3*Position_H+α4*Position_V
wherein, Importance _ Score represents the Importance Score of the person in the picture, 0<Importance_Score≤1；α1、α2、α3、α4Respectively, the weights of the four features.
7. The method for identifying important persons in pictures according to any one of claims 1 to 6, wherein the preset threshold is 0.7.
8. A system for identifying important persons in pictures, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of the preceding claims 1 to 7 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010641879.3A CN111967312B (en) | 2020-07-06 | 2020-07-06 | Method and system for identifying important persons in picture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010641879.3A CN111967312B (en) | 2020-07-06 | 2020-07-06 | Method and system for identifying important persons in picture |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111967312A true CN111967312A (en) | 2020-11-20 |
| CN111967312B CN111967312B (en) | 2023-03-24 |
Family
ID=73361134
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010641879.3A Expired - Fee Related CN111967312B (en) | 2020-07-06 | 2020-07-06 | Method and system for identifying important persons in picture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111967312B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023147764A1 (en) * | 2022-02-04 | 2023-08-10 | Huawei Technologies Co., Ltd. | Systems, methods, and media for main group identification in images via social relation recognition |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108563999A (en) * | 2018-03-19 | 2018-09-21 | 特斯联（北京）科技有限公司 | A kind of piece identity's recognition methods and device towards low quality video image |
| CN109284729A (en) * | 2018-10-08 | 2019-01-29 | 北京影谱科技股份有限公司 | Method, apparatus and medium based on video acquisition human face recognition model training data |
| CN110427795A (en) * | 2019-01-28 | 2019-11-08 | 厦门瑞为信息技术有限公司 | A kind of property analysis method based on head photo, system and computer equipment |
| CN111008558A (en) * | 2019-10-30 | 2020-04-14 | 中山大学 | Picture/video important person detection method combining deep learning and relational modeling |
-
2020
- 2020-07-06 CN CN202010641879.3A patent/CN111967312B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108563999A (en) * | 2018-03-19 | 2018-09-21 | 特斯联（北京）科技有限公司 | A kind of piece identity's recognition methods and device towards low quality video image |
| CN109284729A (en) * | 2018-10-08 | 2019-01-29 | 北京影谱科技股份有限公司 | Method, apparatus and medium based on video acquisition human face recognition model training data |
| CN110427795A (en) * | 2019-01-28 | 2019-11-08 | 厦门瑞为信息技术有限公司 | A kind of property analysis method based on head photo, system and computer equipment |
| CN111008558A (en) * | 2019-10-30 | 2020-04-14 | 中山大学 | Picture/video important person detection method combining deep learning and relational modeling |
Non-Patent Citations (5)
| Title |
|---|
| 张志华等: "基于多元特征的分块人物关系识别系统", 《计算机应用》 * |
| 王维兰等: "基于poselets的特定位置人物多姿势提取", 《计算机系统应用》 * |
| 胥小波等: "基于语义分析的互联网人物信息提取", 《信息安全与通信保密》 * |
| 苏雪平等: "新闻图像中重要人物的自动标志", 《计算机辅助设计与图形学学报》 * |
| 马博宇等: "基于AdaBoost算法的人脸识别系统的研究与实现", 《仪器仪表学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023147764A1 (en) * | 2022-02-04 | 2023-08-10 | Huawei Technologies Co., Ltd. | Systems, methods, and media for main group identification in images via social relation recognition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111967312B (en) | 2023-03-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11495050B2 (en) | Method for distinguishing a real three-dimensional object from a two-dimensional spoof of the real object | |
| Li et al. | Localizing and quantifying damage in social media images | |
| US11416672B2 (en) | Object recognition and tagging based on fusion deep learning models | |
| US9202137B2 (en) | Foreground object detection from multiple images | |
| US7376270B2 (en) | Detecting human faces and detecting red eyes | |
| CN111753767A (en) | Method and device for automatically correcting operation, electronic equipment and storage medium | |
| US20190294921A1 (en) | Field identification in an image using artificial intelligence | |
| US20230004604A1 (en) | Ai-augmented auditing platform including techniques for automated document processing | |
| CN110858327A (en) | Method of validating training data, training system and computer program product | |
| CN110909618A (en) | Pet identity recognition method and device | |
| CN103824090A (en) | Adaptive face low-level feature selection method and face attribute recognition method | |
| US20220207860A1 (en) | Similar area detection device, similar area detection method, and computer program product | |
| US20240087368A1 (en) | Companion animal life management system and method therefor | |
| WO2023088174A1 (en) | Target detection method and apparatus | |
| CN112597909A (en) | Method and equipment for evaluating quality of face picture | |
| CN111967312B (en) | Method and system for identifying important persons in picture | |
| Shehzadi et al. | Towards end-to-end semi-supervised table detection with deformable transformer | |
| US11803585B2 (en) | Method and apparatus for searching for an image and related storage medium | |
| US20140247992A1 (en) | Attribute recognition via visual search | |
| CN109657710B (en) | Data screening method and device, server and storage medium | |
| CN109635798A (en) | A kind of information extracting method and device | |
| Johnson et al. | Using Ensemble Convolutional Neural Network to Detect Deepfakes Using Periocular Data | |
| US11748451B2 (en) | Machine learning techniques for differentiability scoring of digital images | |
| CN111428679B (en) | Image identification method, device and equipment | |
| Hoshino et al. | Inferencing the best AI service using Deep Neural Networks |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20230324 |