CN111639689A - Face data processing method and device and computer readable storage medium - Google Patents
Face data processing method and device and computer readable storage medium Download PDFInfo
- Publication number
- CN111639689A CN111639689A CN202010431025.2A CN202010431025A CN111639689A CN 111639689 A CN111639689 A CN 111639689A CN 202010431025 A CN202010431025 A CN 202010431025A CN 111639689 A CN111639689 A CN 111639689A
- Authority
- CN
- China
- Prior art keywords
- face
- snapshot data
- abnormal
- snapshot
- determining
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
- G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
-
- 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
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Artificial Intelligence (AREA)
- Collating Specific Patterns (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The application discloses a face data processing method, a face data processing device and a computer readable storage medium, wherein the method comprises the following steps: respectively determining the quantity of face snapshot data belonging to the same person aiming at all face snapshot data acquired by the same acquisition device in a preset statistical period; determining abnormal face snapshot data quantity from a plurality of face snapshot data quantities corresponding to the same acquisition device; and determining the face image corresponding to the abnormal face snapshot data quantity as a suspected abnormal face image. According to the method and the device, based on the statistical rule that the snapshot amount of the advertising board is obviously higher than that of other non-advertising board faces, the number of abnormal face images is determined from the number of the plurality of face images acquired by the same acquisition device, and the face images corresponding to the number of the abnormal face images are determined to be suspected abnormal face images, so that preliminary identification of the suspected advertising board is quickly and conveniently realized.
Description
Technical Field
The present application relates to the field of data processing technologies, and in particular, to a method and an apparatus for processing face data, and a computer-readable storage medium.
Background
With the rapid development of big data technology, face snapshot is the most direct and effective way to manage social security in the security industry. The data contains certain social behaviors and social relations of human beings, and the face snapshot data is effectively and intelligently analyzed, so that an extremely effective effect can be brought to the stabilization and public security of the society. However, the face snapshot machine has certain limitations in the snapshot stage and the face recognition analysis stage, and cannot avoid the snapshot of two-dimensional planar face images and the snapshot of other suspected face images; therefore, a waste of storage resources is caused, and the correctness of the analysis function based on the face data is affected.
Therefore, the limitation of the face snapshot machine and the face recognition stage is improved based on data analysis, the storage pressure of face data can be reduced to a certain extent, and the waste of storage resources is avoided; meanwhile, the quality of the data can be improved, so that the accuracy of the analysis function based on the human face data is improved, and the method has important research significance on the application and the structural design of a human face system.
In the prior art, a snapshot device is mainly used for acquiring a face picture, modeling and analyzing a face and endowing a human _ id; and then performing warehousing persistent storage. However, due to the limitation of face recognition of the snapshot device, the back end carries out modeling analysis on the front-end snapshot picture, and accurate recognition on the two-dimensional plane face picture cannot be carried out. The prior art has the following defects:
at the present stage, capturing a face picture through capturing equipment, storing the picture, identifying the face by a face identification analysis algorithm, and marking a human _ id label; however, since the picture is analyzed, the situation that the two-dimensional plane face image is captured cannot be avoided; the snapshot device also has certain limitations, and the snapshot pictures are not all face pictures, and some things suspected of being a face can be snapshot.
Disclosure of Invention
The application aims to provide a face data processing method and device and a computer readable storage medium. The following presents a simplified summary in order to provide a basic understanding of some aspects of the disclosed embodiments. This summary is not an extensive overview and is intended to neither identify key/critical elements nor delineate the scope of such embodiments. Its sole purpose is to present some concepts in a simplified form as a prelude to the more detailed description that is presented later.
According to an aspect of an embodiment of the present application, there is provided a face data processing method, including:
respectively determining the quantity of face snapshot data belonging to the same person aiming at all face snapshot data acquired by the same acquisition device in a preset statistical period;
determining abnormal face snapshot data quantity from a plurality of face snapshot data quantities corresponding to the same acquisition device;
and determining the face image corresponding to the abnormal face snapshot data quantity as a suspected abnormal face image.
According to another aspect of the embodiments of the present application, there is provided a face data processing apparatus, including:
the face snapshot system comprises a first module, a second module and a third module, wherein the first module is used for respectively determining the number of face snapshot data belonging to the same person aiming at all face snapshot data acquired by the same acquisition device in a preset statistical period;
the second module is used for determining the abnormal face snapshot data quantity from the plurality of face snapshot data quantities corresponding to the same acquisition device;
and the third module is used for determining the face image corresponding to the abnormal face snapshot data quantity as a suspected abnormal face image.
According to another aspect of embodiments of the present application, there is provided a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the above-mentioned face data processing method.
According to another aspect of an embodiment of the present application, there is provided an electronic device including: the human face data processing method comprises a memory, a processor and a computer program, wherein the computer program is stored in the memory, and the processor runs the computer program to execute the human face data processing method.
The technical scheme provided by one aspect of the embodiment of the application can have the following beneficial effects:
according to the face data processing method provided by the embodiment of the application, based on the statistical rule that the snapshot amount of the advertising board is obviously higher than that of the other non-advertising board faces, the number of abnormal face images is determined from the number of the plurality of face images acquired by the same acquisition device, and the face images corresponding to the number of the abnormal face images are determined to be suspected abnormal face images, so that preliminary identification of the suspected advertising board is quickly and conveniently realized.
Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the embodiments of the application, or may be learned by the practice of the embodiments. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be 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 described in the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 shows a flow diagram of a face data processing method of an embodiment of the present application;
FIG. 2 shows a flowchart of one implementation of step S10 of the embodiment shown in FIG. 1;
FIG. 3 shows a flowchart of another implementation of step S10 of the embodiment shown in FIG. 1;
fig. 4 is a block diagram showing a configuration of a face data processing apparatus according to another embodiment of the present application;
FIG. 5 shows a block diagram of a face data processing system of an embodiment of the present application;
FIG. 6 shows a flow diagram of a face data processing method of an embodiment of the present application;
fig. 7 shows a schematic diagram of a snapshot device of an embodiment of the present application for snapping a picture of a human face on a billboard.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is further described with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. 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 application.
It will be understood by those within the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
As shown in fig. 1, a first embodiment of the present application provides a face data processing method, including:
and S10, determining a suspected abnormal face image from the face images acquired by the same acquisition device.
In certain embodiments, as shown in fig. 2, step S10 includes:
s20, determining the number of face images belonging to the same person aiming at all face images acquired by the same acquisition device in a preset statistical period; for example, the number N1 of face images of the person a, the number N2 of face images of the person B, the number N3 of face images of the person C, and the like are finally determined by the face image clustering algorithm in all the face images acquired by the acquisition Device 1.
S21, determining the number of abnormal face images from the number of the face images corresponding to the same acquisition device; the above example is adopted, namely the number of abnormal face images is determined from the number of face images of N1, N2, N3 and the like belonging to the same person; for example, when N1 is 500, N2 is 40, and N3 is 50, since N1 is much larger than N2 and N3, it may be determined that N1 is an abnormal number of face images. The number of abnormal face images may also be determined in other manners, which is not limited in this embodiment.
And S22, determining the face images corresponding to the abnormal face image number as suspected abnormal face images. Following the above example, after determining that N1 is the number of abnormal face images, the face images of the N1 persons a can be determined to be suspected abnormal face images.
The acquisition means may be a snap-shot device for snapping a photograph containing an image of a human face, for example a face snap-shot machine. The snapshot device is generally arranged in public places with large pedestrian volume for face snapshot, such as railway stations, and two-dimensional planar face images of non-real persons, such as face photos on billboards, often exist in the public areas. When the snapshot machine detects that the continuous change of the picture exceeds a certain amplitude (a preset threshold amplitude), the snapshot is carried out.
For example, the threshold number of frames is set to 96, and when the number of frames continuously changing is detected by the snapshot machine to exceed 96, the snapshot is performed, and when the number of frames continuously changing does not exceed 96, the snapshot is not performed. Therefore, when a moving object exists in front of the billboard, the snapshot machine detects that the continuous change of the picture exceeds the preset threshold value, and the area of the billboard is snapshot. Because the area of the billboard is larger and the billboard generally has a portrait, the snapshot volume of the billboard is obviously higher than that of the faces of other non-billboards. Based on the statistical rule, the number of abnormal face images is determined according to the number of the plurality of face images acquired by the same acquisition device, and the face images corresponding to the number of the abnormal face images are determined to be suspected abnormal face images, so that the preliminary identification of the suspected billboard is quickly and conveniently realized.
In some embodiments, determining an abnormal face snapshot data amount from a plurality of face snapshot data amounts corresponding to the same acquisition device includes:
determining the total number of face snapshot data acquired by the same acquisition device in the preset statistical period;
aiming at the face snapshot data quantity of any person, if the ratio of the face snapshot data quantity to the total number of the face snapshot data exceeds a first preset threshold value, determining that the face snapshot data quantity is an abnormal face snapshot data quantity; or if the ratio of the number of the face snapshot data to the total number of the face snapshot data with a preset proportion exceeds a second preset threshold, determining that the number of the face snapshot data is an abnormal number of face images. Specifically, the face snapshot data may be a face image.
In some embodiments, determining the number of abnormal facial images from a plurality of the numbers of facial images corresponding to the same acquisition device includes: in the face images acquired by the same acquisition device in a preset statistical period, if the ratio of the number of the face images of a certain person to the number of all the face images acquired by the acquisition device in the preset statistical period exceeds a first preset threshold or the ratio of the number of the face images of the certain person to the number of the face images of the preset proportion of the number of all the face images exceeds a second preset threshold, the number of the face images of the person is the number of abnormal face images.
Namely: in the face images collected by the same collecting device in a preset statistical period, if the ratio of the number a of the face images of a certain person to the total number B of the face images collected by the collecting device exceeds a first preset threshold D, or the ratio of the number a of the face images of a certain person to the total number B of the face images collected by the collecting device in a preset ratio (the specific ratio can be set according to actual needs), the ratio in the number C is recorded as exceeding a second preset threshold E, and the face images of the person are suspected abnormal face images. Namely, if A/B is larger than D or A/C is larger than E, the face image of the person is a suspected abnormal face image. The snapshot amount of the advertising board is obviously higher than that of the faces of other non-advertising boards, so that the ratio of the snapshot amount of the advertising board to the total or partial number of the face images acquired by the acquisition device is larger, when the ratio exceeds a preset threshold value, the number of the face images is determined to be the number of abnormal face images, and the face images corresponding to the abnormal face images are suspected abnormal face images such as suspected advertising board face images.
In some embodiments, determining the number of abnormal facial images from a plurality of the numbers of facial images corresponding to the same acquisition device includes:
selecting the number of the first n face images from the number of the plurality of face images corresponding to the same acquisition device according to the descending order of the number of the face snapshot data; the proportion of the sum of the number of the first n face images in the number of all the face images acquired by the same acquisition device exceeds a third preset threshold;
and if the proportion of the number of the face images of a certain person in the sum of the number of the first n face images exceeds a fourth preset threshold value, determining that the number of the face images is the abnormal number of the face images.
In some embodiments, the face image acquired by the same acquisition device, the acquisition time of each face image and the acquisition device identifier of the acquisition device form a face image set corresponding to the acquisition device, and then suspected abnormal face images are screened from the face image set.
In certain embodiments, as shown in fig. 3, step S10 includes:
S101、
and marking all the face images belonging to the same face by using unique face identifications aiming at all the face images acquired by the same acquisition device in a preset statistical period, and calculating the number of the face images corresponding to each face identification.
All the face images of the same face are marked by the same face identification, and the face images of different faces are respectively marked by different face identifications. Each face corresponds to a face identifier.
For example, the face identification may be human1, human2, human3 … …, etc., and the number of face images identified as human1, the number of face images identified as human2, the number of face images identified as human3, etc. are calculated.
S102, selecting comparison marks from all the face marks, and representing the marked face images to form a comparison set by using the comparison marks, wherein in some embodiments, the ratio of the sum of the number of the face images of all the comparison marks to the total number of all the face images acquired by the same acquisition device reaches or exceeds a first preset threshold; the comparison identification is the face identification with the marked face image quantity ranking belonging to a large number of preset values; the preset numerical value is determined according to the preset first threshold value and the total number of all the face images.
When the ratio of the number of face images in the comparison set to the number of face images acquired by the acquisition device is set to 100%, all the face images acquired by the acquisition device need to form the comparison set. In order to reduce the amount of calculation, the threshold value is generally set to a value less than 100%.
S103, determining the ratio of the number of the face images marked by each contrast identifier to the total number of the face images in the contrast set.
For example, as shown in table 1, the number of face images corresponding to 7 face identifiers captured by four capturing devices is shown.
Table 1 snapshot statistics top7 for snapshot device
| Snapshot equipment identification | human1 | human2 | human3 | human4 | human5 | human6 | human7 |
| X33010853011210000150 | 776 | 52 | 9 | 6 | 6 | 6 | 4 |
| X33010351001210000022 | 339 | 51 | 10 | 6 | 6 | 6 | 6 |
| X33010462001210010427 | 2536 | 16 | 16 | 10 | 10 | 9 | 9 |
| X33010456001210010385 | 2306 | 19 | 5 | 4 | 4 | 3 | 3 |
The human1, the human2, the human3, the human4, the human5, the human6 and the human7 are face identifications of 7 persons, and are respectively used for identifying corresponding face images. Each face identifier uniquely identifies a face. The faces corresponding to the 7 face identifications are seven faces which are respectively captured by the acquisition device for the most times. The ratio of the face images of the people with the corresponding image quantity ranking belonging to the top7 to the number of the face images collected by the same collecting device reaches 90 percent. In this embodiment, the face images of the top7 people with the corresponding image quantity ranking are selected, and the preset value is 7 in this embodiment.
X33010853011210000150, X33010351001210000022, X33010462001210010427 and X33010456001210010385 are the identifications of the acquisition devices.
Setting:
and N is the snapshot amount of the device of the snapshot device.
P is the snapshot amount of the billboard which is snapshot by the device.
And R is the snapshot amount of the normal pedestrian which is snapshot by the device.
Then there is: P/N > > R/N.
Taking human1 as an example, the number of human face images of human1 captured by X33010853011210000150 is 776, the number of human face images of human1 captured by X330103510351001210000022 is 339, the number of human face images of human1 captured by X33010462001210010427 is 2536, and the number of human face images of human1 captured by X33010456001210385 is 2306.
A percentage threshold is set, for example, 30% may be set, and 30% may be set because a billboard generally does not exceed 3 facial pictures. Therefore, if the billboard is used, the snapshot occupation ratio can exceed 30% with a high probability. When the proportion of the snapshot amount of the same face identifier acquired by the same acquisition device in the total number of the face images of the 7 persons acquired by the acquisition device exceeds the proportion threshold value, the face identifier is more likely to be the billboard face. All the face images corresponding to the face identifier are suspected abnormal face images meeting preset conditions, where the preset conditions are that the ratio of the number of face images corresponding to the face identifier in the total number of face images of the 7 persons acquired by the acquisition device exceeds a set ratio threshold (here, 30%).
By the above statistical rules, it can be calculated from the data in table 1 that the number of face images of human1 captured by X33010853011210000150 is 776, the number of face images of human7 is 4, the number of face images of these 7 faces captured is 859, 776/859 ≈ 90.34%, 4/859 ≈ 0.466%, the proportion of the face images of human1 in the first 7 names with the highest captured amount exceeds 90%, and the face images of human1 are suspected abnormal face images;
the number of face images of human1 captured by X3301035103510351001210000022 is 339, the number of face images of the 7 faces captured is 424, 339/424 is about 79.95%, the proportion of the face images of human1 in the first 7 names with the highest capturing amount exceeds 70%, and the face images of human1 are suspected face images if the proportion is the highest; the number of face images of human1 captured by X33010462001210010427 is 2536, the number of face images of the 7 faces captured is 2606, 2536/2606 ≈ 97.31%, the proportion of the number of face images of the face identification human1 exceeds 90%, and thus the face identification human1 is recognized as an abnormal human _ id (i.e., abnormal face identification); the number of face images of human1 captured by X33010456001210385 is 2306, the number of face images of the 7 faces captured is 2344, 2306/2344 is approximately 98.38%, and the proportion of the face images of human1 in the first 7 names with the highest capturing amount exceeds 90%, so that the face image of human1 is a suspected abnormal face image.
And S104, determining the contrast identification corresponding to the ratio larger than the preset ratio threshold value as a suspected abnormal face identification, wherein the face image marked by the suspected abnormal face identification is a suspected abnormal face image.
The face images acquired by the four acquisition devices have the face identification man1 with the number ratio of the face images exceeding 30%, so that the face identification man1 is recognized as a suspected abnormal human _ id (i.e., a suspected abnormal face identification). By using the suspected abnormal face identification obtained in the above steps as a reference, the suspected abnormal face identification can be used for judging whether the subsequently obtained face image is an abnormal face image or not under the condition of low requirement on the accuracy of the abnormal face identification. In the process of acquiring the face image in real time, the face image which is detected by the suspected abnormal face image marked by the suspected abnormal face identification can be used for filtering the acquired face image.
However, it cannot be determined that the face image corresponding to human1 is the face image on the billboard by one hundred percent, for example: when a person stands under the snapshot machine and does not move, the situation is normal snapshot. If the requirement on the identification accuracy of the abnormal face image is high, further analysis is needed according to a space-time rule. Therefore, in another embodiment, the face data processing method further includes:
determining a suspected abnormal face image set formed by face images belonging to the same person in the suspected abnormal face images acquired by the acquisition devices; for example, the face images marked as the same face identifier and acquired by the acquisition devices may be acquired, that is, the face images belonging to the same person may be acquired to form a suspected abnormal face image set.
If the time-space information of at least two face images in the suspected abnormal face image set meets an abnormal condition, determining the face images in the suspected abnormal face image set as abnormal face images; the spatiotemporal information comprises the acquisition time and the acquisition position of the face image.
The space-time information meets abnormal conditions, such as the acquisition time difference is smaller than a preset time interval, and the distance of the acquisition position is larger than a preset distance; alternatively, the ratio of the distance between the acquisition positions and the acquisition time difference exceeds a preset speed threshold, and the like, which is not limited herein.
Specifically, as shown in fig. 1, in some embodiments, the face data processing method further includes the following steps:
s30, determining a suspected abnormal face snapshot data set formed by face snapshot data belonging to the same person in the suspected abnormal face snapshot data acquired by each acquisition device in the preset statistical period; if the time-space information of at least two pieces of snapshot data in the suspected abnormal face snapshot data set meets an abnormal condition, determining the face image in the suspected abnormal face snapshot data set as an abnormal face image; the time-space information comprises the snapshot time of the face snapshot data and a snapshot device.
In some embodiments, before determining that a face image in the suspected abnormal face snapshot dataset is an abnormal face image if the spatiotemporal information of at least two pieces of snapshot data in the suspected abnormal face snapshot dataset satisfies an abnormal condition, the method further includes:
sequencing all the snapping devices according to the earliest snapping time;
and deleting the face snapshot data of each snapshot device from the earliest snapshot time of each snapshot device to the earliest snapshot time of the next snapshot device in the sequence.
In some embodiments, if the spatiotemporal information of at least two pieces of snapshot data in the suspected abnormal face snapshot dataset satisfies an abnormal condition, determining that the face image in the suspected abnormal face snapshot dataset is an abnormal face image, including:
determining at least two pieces of snapshot data which are different in snapshot device and have a snapshot time difference smaller than a preset time threshold;
if the distance between the snapshot devices corresponding to the at least two pieces of snapshot data is larger than a preset distance threshold, determining that the time-space information of the snapshot data in the suspected abnormal face snapshot data set meets an abnormal condition;
and determining the face image in the suspected abnormal face snapshot data set as an abnormal face image.
A face data processing method according to a second embodiment of the present application includes the following steps:
1) and determining a suspected abnormal face image from the face images acquired by the same acquisition device.
2) And forming a face image data set by using the suspected abnormal face images acquired by the acquisition devices, the acquisition time of each suspected abnormal face image, the identification of the acquisition devices and the identification of the suspected abnormal face.
For example, the face image (suspected abnormal face image) corresponding to human1, the acquisition time of the face image acquired by human1, and the identifications X33010853011210000150, X33010351001210000022, X33010462001210010427, and four acquisition devices
X33010456001210010385 constitutes a face image dataset.
3) Determining the acquisition distance and the acquisition time difference corresponding to any two suspected abnormal face images marked by the same suspected abnormal face identification in the face image data set; the acquisition distance refers to the distance between acquisition devices for acquiring two suspected abnormal face images; the acquisition time difference refers to the time difference between the acquisition times of the two suspected abnormal face images.
4) And judging whether the same suspected abnormal face identification is an abnormal face identification or not according to the acquisition distance and the acquisition time difference.
In some embodiments, determining whether the same suspected abnormal face identifier is an abnormal face identifier according to the acquisition distance and the acquisition time difference includes:
(1) determining a ratio of the acquisition distance to the acquisition time difference;
(2) and judging whether the suspected abnormal face identification is abnormal face identification or not according to the ratio.
Step (2) judging whether the suspected abnormal face identification is an abnormal face identification according to the ratio, comprising the following steps:
comparing each ratio with a preset threshold, and if the ratio is greater than the preset threshold, identifying the suspected abnormal face identifier corresponding to the ratio as an abnormal face identifier; otherwise, the face is not abnormal face identification.
For example, suppose that two face images of the same face are at time t respectively3And a point in time t4Quilt device D3And device D4Capture the time differenceIs Δ t ═ t3-t4L, collecting distance D as device D3And D4The ratio v of the acquisition distance to the acquisition time difference is Δ t/d. This ratio represents the speed of movement of the person. If this speed exceeds the trial value, an anomaly is present. For example, the highest speed limit in all the snapshot ranges is 70km/h, i.e. the preset threshold is 70km/h, if the calculated speed is 1000km/h, 1000km/h>70km/h, the face image on the captured billboard can be determined.
In some embodiments, determining whether the same suspected abnormal face identifier is an abnormal face identifier according to the acquisition distance and the acquisition time difference includes:
judging whether the acquisition distance and the acquisition time difference both meet respective corresponding preset threshold conditions; if yes, the suspected abnormal face identification is abnormal face identification, and if not, the suspected abnormal face identification is not abnormal face identification.
By spatiotemporal regularity is meant the existence of some regularity in the longitudinal alignment in both the temporal and spatial dimensions. A time period threshold and a distance threshold are preset. If the interval of the two capturing devices to the capturing time points of the same billboard is smaller than the preset time period threshold, the capturing time points of the two capturing devices belong to the same capturing time sequence. If the distance between the two pieces of snapshot equipment for snapshotting the same billboard is larger than a preset distance threshold value, the two pieces of snapshot equipment belong to the same snapshot space sequence. If two face images of a certain person belong to the same snapshot time sequence and the same snapshot space sequence, the two face images can be considered as abnormal face images.
Example 1: the same face picture a on the billboard is captured by different devices with a distance of 17km at adjacent time or the same time and stored in a database, as shown in fig. 7.
Table 2 certain face identification humanId is snapshotted to record TOP5
Table 2 shows how a certain humanId is snapshotted at different times; such as: at T1The time instants are captured by X33011067001210001382, X33010454001210010372 and X33010456001210010385 at the same time, and denoted by 1.
Setting:
t1,t2∈T1i.e. t1、t2Belonging to the same snapshot time sequence (i.e. a set of time points having a length less than a certain length, e.g. T ═ T [ T ])1,t2]And T represents T1And t2The set of all time instants in between);
D1and D2Respectively numbering two pieces of snapshot equipment;
t1is D1Time to snap to the billboard;
t2is D2Time to snap to the billboard;
then there is:
acquisition time difference | t1-t2|<threshold (default 2min)
And collects the distance D (D)1,D2)>threshold (default 2 km).
For example, the preset threshold is set to 400km/h, if the ratio v of the acquisition distance to the acquisition time difference is greater than 400km/h, the face in the face image can be determined to be an abnormal face, and the face image is a face image on a billboard; otherwise, the face is not an abnormal face. The judgment method actually judges according to the speed which can be reached by a real person. For example, the speed of a person riding a high-speed rail can reach 380km/h at most, and if the ratio v exceeds 400km/h, the possibility of an abnormal face is very high. The preset threshold value can be adjusted according to actual conditions, and is generally set as the highest speed limit of the snapshot coverage area.
And acquiring face identification/identity information of the abnormal face image aiming at the abnormal face, and filtering the abnormal face data according to the face identification/identity information in the process of acquiring the face image snapshot data in real time.
For example, if the suspected abnormal face image corresponding to the same humanId corresponds to an abnormal face (for example, a face image on a billboard), the face image corresponding to the humanId 1 and the corresponding acquisition device identifier and acquisition time are directly filtered in the process of acquiring the face image in real time.
In some embodiments, the abnormal face identifier in the face image data set, the abnormal face image, the acquisition device identifier of the abnormal face image, and the acquisition time are additionally stored in other databases to be used as comparison information, when a new face image is acquired, the new face image is compared with the abnormal face image, if the new face image is the face image of the same person as the abnormal face image, the face corresponding to the new face image is also the abnormal face, and the new face image is not stored in the database of the normal face image.
A third embodiment of the present application provides a face data processing method, based on the first embodiment, before the step S10 of screening a suspected abnormal face image from face images acquired by the same acquisition device, the method further includes:
s00, extracting all face images from all photos captured by all the acquisition devices;
the captured picture is captured by a capturing device such as a capturing device; carrying out face recognition on the captured photos, and recognizing faces from the photos;
comparing the extracted face images, assigning the same face identification to the face images belonging to the same person, for example, comparing the extracted face images, and assigning the same face identification human _ id to the identified face images belonging to the same person, namely, each face corresponds to a unique face identification, and all the face images of the face are marked by the face identification;
and S01, acquiring the identification and acquisition time of the acquisition device for acquiring each face image.
And searching the equipment identification of the snapshot equipment for snapshotting each face image and the snapshot time for snapshotting the face image.
And the face images of the same person are marked by one face identification, and the identification and the snapshot time of the corresponding snapshot device of each face image are associated with the face image, so that the face image is convenient to search.
On the basis of the second embodiment of the present application, as shown in fig. 2, before determining the acquisition distance and the acquisition time difference corresponding to any two suspected abnormal face images marked by the same suspected abnormal face identifier in the face image data set in step 3), the method further includes: 2') performing data cleaning on the suspected abnormal face images acquired by the acquisition devices to obtain a cleaned face image data set. In this embodiment, in step 3), an acquisition distance and an acquisition time difference corresponding to any two suspected abnormal face images marked by the same suspected abnormal face identifier in the washed face image dataset are determined. Through data cleaning, repeated face images which are captured in the same time period and the same place in the face image data set are removed, the data processing workload is reduced, the data processing efficiency is improved, and the data processing pressure is reduced. The data cleaning operation is to delete the acquisition record of each acquisition device between the earliest acquisition time of each acquisition device and the earliest acquisition time of the next acquisition device.
In some embodiments, the step 2') of performing data cleaning on the suspected abnormal face image acquired by each of the acquisition devices to obtain a cleaned face image dataset includes:
s001, sequencing the identifications of the acquisition devices according to the sequence of the earliest acquisition time; the earliest acquisition time is the acquisition time when an acquisition device acquires the suspected abnormal face image earliest;
s002, removing the suspected abnormal face images collected in the corresponding adjacent time period by each collecting device identifier in the sequence to obtain a cleaned face image data set consisting of the remaining suspected abnormal face images; the contiguous time period is the time period between two earliest acquisition times that are adjacent in the ordering.
The range of the visual domain of the snapshot device is limited, therefore, if the same person does not leave the visual domain within a certain time period, a plurality of photos can be snapshot, if all face images of the person in the photos are stored, a large amount of storage space can be occupied, and the waste of the storage space is caused.
The purpose of the snapshot face image is, for example, tracking a specific person, and the like, and by performing processing statistical analysis on the snapshot face image, it is helpful to find the specific person and determine that the specific person is at a certain place at a certain time. For example, if a specific human nail is in the visual field of the capturing device a at 8 o ' clock 10 min 12 sec to 8 o ' clock 40 min 22 sec on a certain day, and is captured by a to a plurality of photos, and is in the visual field of the capturing device B at 8 o ' clock 41 min 41 sec to 9 o ' clock 5 min 53 sec, and is captured by B to a plurality of photos, and then returns to the visual field of a at 9 o ' clock 6 min 17 sec to 9 o ' clock 24 min 49 sec, then when sorting is performed according to the order of the earliest acquisition time, a is located before B and adjacent to B, the earliest acquisition time of a is 8 o ' clock 10 min 12 sec, and the earliest acquisition time of B is 8 o ' clock 41 min 41 sec, the adjacent time period corresponding to a is a time period between 8 o ' clock 10 min 12 sec to 8 o ' clock 41 min 41 sec, and when data cleaning is performed, the face image of the nail acquired by a in the time period between 8 o ' clock 10 min 12 sec to 8 o ' clock 41 min 41 sec is deleted, and only the face image captured at 8 o ' clock 10 min 12 sec is kept, or if a certain face image in the adjacent time period is shot more clearly than the face image shot at the earliest time, the clearer face image in the adjacent time period can be reserved, and the face image shot at the earliest time and other face images in the adjacent time period can be removed. Therefore, the data required to be stored is greatly reduced, the storage of repeated face images is avoided, the storage space is saved, and the data processing efficiency is improved.
In some embodiments, the step 2') of performing data cleaning on the suspected abnormal face image acquired by each of the acquisition devices to obtain a cleaned face image dataset includes:
s1, finding out first acquisition time and a first acquisition device identifier from the suspected abnormal face image dataset; the first acquisition time is the earliest acquisition time for acquiring a suspected abnormal face image; the first acquisition device identifier is an acquisition device identifier for acquiring the suspected abnormal face image at the first acquisition time;
s2, finding out second acquisition time and a second acquisition device from the suspected abnormal face image data set; the second acquisition time is the acquisition time when the acquisition devices except the first acquisition device acquire the suspected abnormal face image earliest; the second acquisition device identifier is an acquisition device identifier for acquiring the suspected abnormal face image at the second acquisition time;
s3, removing suspected abnormal face images acquired by the first acquisition device in a first time period; the first time period is a time period between the first acquisition time and the second acquisition time;
and S4, using the second acquisition time as a new first acquisition time, using the second acquisition device as a new first acquisition device, and turning to the step S2 until the second acquisition time is not searched, and ending to obtain the cleaned face image data set.
In some embodiments, for reading data speed, calculation speed, and the like, the acquisition records of the capturing device are selected as the calculation data structure:
i. and taking a spatial point (i.e. a snapshot device) as a keyword (key).
The value is the time sequence in which the human _ id is captured by the capturing device (key).
The number of capturing devices capturing to a certain human _ id in a day is limited, and the length of the formed spatial sequence is far shorter than the length of the time sequence formed by capturing time.
A certain human _ id is repeatedly and uninterruptedly captured by a capturing device, so that a capturing time sequence of a single capturing device is overstaffed, and only one capturing record of a certain device is ensured to exist in a period of time through data cleaning until the capturing is carried out by different devices.
The data structure is set as:
[d1,<t11,t12,t13……t1x1>]
[d2,<t21,t22,t23……t2x2>]
……
[dm,<tm1,tm2,tm3……tmxn>]
for example, the cleaning process includes the following steps:
1) finding the minimum time (t) from all the snapshot time seriesmin) And recording tminAssociated device (device)key)；
2) Slave non-devicekeyIn other sequences of devices, find out that is greater than tminMinimum time (t) oftemp) And recording the associated device (device)temp)；
3) Removing devicekeyIs greater than t in the snapshot time sequenceminLess than ttempAll time points of (a);
4) will ttempIs assigned to tmin，devicetempAssign value to devicekeyThe next cycle is performed.
5) Until ttempIf not, the loop is ended.
The face data processing method provided by the embodiment can accurately identify and remove the abnormal face image, improves the efficiency of acquiring the face data, avoids the waste of storage space, reduces the storage pressure, and improves the acquisition efficiency of effective face data.
In some embodiments, if the spatiotemporal information of at least two pieces of snapshot data in the suspected abnormal face snapshot dataset satisfies an abnormal condition, determining that the face image in the suspected abnormal face snapshot dataset is an abnormal face image, including:
determining at least two pieces of snapshot data which are different in snapshot device and have a snapshot time difference smaller than a preset time threshold;
if the distance between the snapshot devices corresponding to the at least two pieces of snapshot data is larger than a preset distance threshold, determining that the time-space information of the snapshot data in the suspected abnormal face snapshot data set meets an abnormal condition;
and determining the face image in the suspected abnormal face snapshot data set as an abnormal face image.
For example,
1) and taking time as Key, and merging the items of which the difference value of the keys is smaller than a snapshot time interval threshold (equal to the time interval threshold in the space-time law).
2) The value of Key is found to have items of different snapshot devices.
And calculating the distances of different capturing devices in a single item, and if the distance is greater than a threshold (equal to a distance interval threshold in a space-time rule), judging that the advertisement board is suspected to be captured.
As shown in fig. 4, another embodiment of the present application further provides a face data processing apparatus, including:
the face snapshot system comprises a first module, a second module and a third module, wherein the first module is used for respectively determining the number of face snapshot data belonging to the same person aiming at all face snapshot data acquired by the same acquisition device in a preset statistical period;
the second module is used for determining the abnormal face snapshot data quantity from the plurality of face snapshot data quantities corresponding to the same acquisition device;
and the third module is used for determining the face image corresponding to the abnormal face snapshot data quantity as a suspected abnormal face image.
In certain embodiments, the apparatus further comprises:
the fourth module is used for determining a suspected abnormal face snapshot data set formed by face snapshot data belonging to the same person in the suspected abnormal face snapshot data acquired by each acquisition device in the preset statistical period; if the time-space information of at least two pieces of snapshot data in the suspected abnormal face snapshot data set meets an abnormal condition, determining the face image in the suspected abnormal face snapshot data set as an abnormal face image; the time-space information comprises the snapshot time of the face snapshot data and a snapshot device.
Another embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the above-mentioned face data processing method.
Another embodiment of the present application further provides an electronic device, including: the human face data processing method comprises a memory, a processor and a computer program, wherein the computer program is stored in the memory, and the processor runs the computer program to execute the human face data processing method.
In another embodiment of the present application, the face data is collected and processed by a face data processing system. As shown in fig. 5, the face data processing system includes the following parts:
1) the acquisition device is used for capturing a face image;
2) a static comparison server: comparing and modeling data captured by consumption front-end capturing equipment (namely an acquisition device), and endowing a unique human face identification human _ id for the recognized human face;
3) a face acquisition module: after the static comparison server finishes face recognition, the data are sent back to Kafka; the face acquisition module consumes the data in the Kafka, and filters the data by performing human _ id comparison; kafka is a high throughput distributed publish-subscribe messaging system;
4) suspected billboard analysis module: and analyzing historical data in the database to identify the suspected human _ id of the billboard and update the suspected human _ id to a filtering cache of the face acquisition module.
Fig. 6 is a flowchart of the face data processing method of the present embodiment. Suspected billboard analysis module analyzes and processes the snapshot data, including:
1. snap feature analysis
The face suspected billboard snapshot data has certain specific statistical rules and space-time rules, and whether the face suspected billboard snapshot data is the billboard data can be accurately identified through the rules.
a) Statistical law
The statistical rule is generated by the snapshot rule of the face snapshot machine. When the snapshot machine detects that the continuous change of the picture exceeds a certain amplitude (a preset threshold amplitude), the snapshot is carried out. For example, the threshold number of frames is set to 96, and when the number of frames continuously changing is detected by the snapshot machine to exceed 96, the snapshot is performed, and when the number of frames continuously changing does not exceed 96, the snapshot is not performed.
Therefore, when a moving object exists in front of the billboard, the snapshot machine detects that the continuous change of the picture exceeds the preset threshold value, and the area of the billboard is snapshot. Because the area of bill-board is bigger, and there is the portrait on the bill-board generally, therefore the snapshot volume of this bill-board is bigger in the proportion of this snapshot device snapshot volume. And carrying out snapshot analysis on the suspected billboard in the first step based on the statistical rule. Reference may be made to the data in table 1 above.
A proportion threshold value is set, for example, it may be set to 30%, and when the proportion of the snapshot amount of a face identifier exceeds the proportion threshold value, the face identifier is more likely to be a billboard face. By the above statistical rules, it is possible to identify human1 as an abnormal human id, but it is not certain that human1 is a billboard, such as: when a person stands under the snapshot machine and does not move, the situation is normal snapshot. Further analysis according to spatiotemporal rules is required.
b) Space-time law
The term "temporal-spatial law" refers to the existence of some laws in the longitudinal alignment in both the temporal and spatial dimensions.
A time period threshold and a distance threshold are preset. If the interval of the two capturing devices to the capturing time points of the same billboard is smaller than the preset time period threshold, the capturing time points of the two capturing devices belong to the same capturing time sequence. If the distance between the two pieces of snapshot equipment for snapshotting the same billboard is larger than a preset distance threshold value, the two pieces of snapshot equipment belong to the same snapshot space sequence. Reference may be made to the examples in table 2 above.
2. Construction of data structures
In order to read data speed, calculation speed and other factors, the device acquisition record is selected as a calculation data structure:
i. and taking a spatial point (i.e. a snapshot device) as a keyword (key).
The value is the time sequence in which the human _ id is captured by the capturing device (key).
The number of capturing devices capturing to a certain human _ id in a day is limited, and the length of the formed spatial sequence is far shorter than the length of the time sequence formed by capturing time.
A certain human _ id is repeatedly and uninterruptedly captured by a capturing device, so that a capturing time sequence of a single capturing device is overstaffed, and data is cleaned, so that only one capturing record of a certain device exists in a period of time until the capturing record is captured by different devices.
Data structure:
[d1,<t11,t12,t13……t1x1>]
[d2,<t21,t22,t23……t2x2>]
……
[dm,<tm1,tm2,tm3……tmxn>]
and (3) cleaning rules:
1) finding the minimum time (t) from all the snapshot time seriesmin) And recording tminAssociated device (device)key)；
2) Slave non-devicekeyIn other sequences of devices, find out that is greater than tminMinimum time (t) oftemp) And recording the associated device (device)temp)；
3) Removing devicekeyIs greater than t in the snapshot time sequenceminLess than ttempAll time points of (a);
4) will ttempIs assigned to tmin，devicetempAssign value to devicekeyThe next cycle is performed.
5) Until ttempIf not, the loop is ended.
Pseudo code example:
1)Begin
2)devicekey,tmin<-Function min(T<device>)
3)WHILE devicetemp,ttemp<-Function min(T<deviceother>and>tmin)≠null
4)DO
5)remove(tmin,ttemp]from T<devicemin>
6)tmin<-ttemp
7)devicekey<-devicetemp
8)DONE
9)END
suspected billboard analysis and calculation:
firstly, carrying out statistical probability analysis based on snapshot equipment, and screening abnormal human _ id collected by the snapshot equipment;
and then, performing space-time analysis on the snapshot equipment corresponding to the abnormal human _ id.
The distribution of the snapshot devices can have a decisive influence on the characteristics of the snapshot data, so that no uniform calculation standard can be given, and only the threshold value can be set according to the situation of the field.
Based on the above calculation data structure, a face snapshot suspected billboard recognition algorithm is provided.
The suspected billboard face image recognition algorithm comprises the following steps:
1) selecting snapshot equipment to be analyzed, and performing grouping statistics on the human _ id collected by the snapshot equipment;
2) when the snapshot amount of the human _ id exceeds a certain threshold (default 30%) of the total snapshot amount of the equipment, marking the human _ id as abnormal human _ id, and acquiring the acquisition records of other snapshot equipment of the human _ id to form a data structure to be analyzed;
3) data cleaning is carried out on a data structure to be analyzed, transposition is carried out, time is taken as a key (keyword), and items with the difference value of the key being smaller than a snapshot time interval threshold (equal to the time interval threshold in the space-time law) are merged;
4) finding out items of which the value of Key has different snapshot devices;
5) and calculating the distances of different capturing devices in a single item, and if the distance is greater than a threshold (equal to a distance interval threshold in a space-time rule), judging that the advertisement board is suspected to be captured.
In the method of the embodiment, in the flow of face image acquisition, the analyzed data of the suspected billboard face image is accurately identified and removed, and the data of the suspected billboard is directly forbidden to be put into a warehouse in the acquisition stage. The data acquisition efficiency can be improved while the pressure of the storage system is reduced.
The method of the embodiment can accurately identify the suspected billboard snapshot record aiming at the analysis of the suspected billboard by face snapshot, can reduce the pressure of a storage system of face data, and simultaneously improves the accuracy of the function of data analysis based on the face data.
It should be noted that:
the term "module" is not intended to be limited to a particular physical form. Depending on the particular application, a module may be implemented as hardware, firmware, software, and/or combinations thereof. Furthermore, different modules may share common components or even be implemented by the same component. There may or may not be clear boundaries between the various modules.
The algorithms and displays presented herein are not inherently related to any particular computer, virtual machine, or other apparatus. Various general purpose devices may be used with the teachings herein. The required structure for constructing such a device will be apparent from the description above. In addition, this application is not directed to any particular programming language. It will be appreciated that a variety of programming languages may be used to implement the teachings of the present application as described herein, and any descriptions of specific languages are provided above to disclose the best modes of the present application.
Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features included in other embodiments, rather than other features, combinations of features of different embodiments are meant to be within the scope of the application and form different embodiments. For example, in the following claims, any of the claimed embodiments may be used in any combination.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The above-mentioned embodiments only express the embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.
Claims (10)
1. A face data processing method is characterized by comprising the following steps:
respectively determining the quantity of face snapshot data belonging to the same person aiming at all face snapshot data acquired by the same acquisition device in a preset statistical period;
determining abnormal face snapshot data quantity from a plurality of face snapshot data quantities corresponding to the same acquisition device;
and determining the face image corresponding to the abnormal face snapshot data quantity as a suspected abnormal face image.
2. The method according to claim 1, wherein the determining of the abnormal number of face snapshot data from the plurality of numbers of face snapshot data corresponding to the same acquisition device comprises:
determining the total number of face snapshot data acquired by the same acquisition device in the preset statistical period;
aiming at the face snapshot data quantity of any person, if the ratio of the face snapshot data quantity to the total number of the face snapshot data exceeds a first preset threshold value, determining that the face snapshot data quantity is an abnormal face snapshot data quantity; or if the ratio of the number of the face snapshot data to the total number of the face snapshot data with a preset proportion exceeds a second preset threshold, determining that the number of the face snapshot data is an abnormal number of face images.
3. The method according to claim 1, wherein the determining of the abnormal number of face snapshot data from the plurality of numbers of face snapshot data corresponding to the same acquisition device comprises:
selecting the number of the front n face snapshot data from the number of the face snapshot data corresponding to the same acquisition device according to the descending order of the number of the face snapshot data; the proportion of the sum of the number of the first n pieces of face snapshot data in the total number of the face snapshot data acquired by the same acquisition device in the preset statistical period exceeds a third preset threshold;
and if the ratio of the number of face snapshot data of a certain person to the sum of the number of the previous n face snapshot data exceeds a fourth preset threshold, determining that the number of the face snapshot data is the abnormal number of the face snapshot data.
4. The method according to any one of claims 1 to 3, further comprising:
determining a suspected abnormal face snapshot data set formed by face snapshot data belonging to the same person in the suspected abnormal face snapshot data acquired by each acquisition device in the preset statistical period; if the time-space information of at least two pieces of snapshot data in the suspected abnormal face snapshot data set meets an abnormal condition, determining the face image in the suspected abnormal face snapshot data set as an abnormal face image; the time-space information comprises the snapshot time of the face snapshot data and a snapshot device.
5. The method according to claim 4, wherein if the spatiotemporal information of at least two pieces of snapshot data in the suspected abnormal face snapshot dataset satisfies an abnormal condition, before determining that the face image in the suspected abnormal face snapshot dataset is an abnormal face image, the method further comprises:
sequencing all the snapping devices according to the earliest snapping time;
and deleting the face snapshot data of each snapshot device from the earliest snapshot time of each snapshot device to the earliest snapshot time of the next snapshot device in the sequence.
6. The method according to claim 4, wherein if the spatiotemporal information of at least two pieces of snapshot data in the suspected abnormal face snapshot dataset satisfies an abnormal condition, determining that the face image in the suspected abnormal face snapshot dataset is an abnormal face image, includes:
determining at least two pieces of snapshot data which are different in snapshot device and have a snapshot time difference smaller than a preset time threshold;
if the distance between the snapshot devices corresponding to the at least two pieces of snapshot data is larger than a preset distance threshold, determining that the time-space information of the snapshot data in the suspected abnormal face snapshot data set meets an abnormal condition;
and determining the face image in the suspected abnormal face snapshot data set as an abnormal face image.
7. A face data processing apparatus, comprising:
the face snapshot system comprises a first module, a second module and a third module, wherein the first module is used for respectively determining the number of face snapshot data belonging to the same person aiming at all face snapshot data acquired by the same acquisition device in a preset statistical period;
the second module is used for determining the abnormal face snapshot data quantity from the plurality of face snapshot data quantities corresponding to the same acquisition device;
and the third module is used for determining the face image corresponding to the abnormal face snapshot data quantity as a suspected abnormal face image.
8. The apparatus of claim 7, further comprising:
the fourth module is used for determining a suspected abnormal face snapshot data set formed by face snapshot data belonging to the same person in the suspected abnormal face snapshot data acquired by each acquisition device in the preset statistical period; if the time-space information of at least two pieces of snapshot data in the suspected abnormal face snapshot data set meets an abnormal condition, determining the face image in the suspected abnormal face snapshot data set as an abnormal face image; the time-space information comprises the snapshot time of the face snapshot data and a snapshot device.
9. A computer-readable storage medium on which a computer program is stored, the program being executed by a processor to implement the face data processing method according to any one of claims 1 to 7.
10. An electronic device, comprising: a memory, a processor and a computer program, the computer program being stored in the memory, the processor running the computer program to perform the face data processing method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010431025.2A CN111639689B (en) | 2020-05-20 | 2020-05-20 | Face data processing method and device and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010431025.2A CN111639689B (en) | 2020-05-20 | 2020-05-20 | Face data processing method and device and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111639689A true CN111639689A (en) | 2020-09-08 |
| CN111639689B CN111639689B (en) | 2023-07-25 |
Family
ID=72331483
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010431025.2A Active CN111639689B (en) | 2020-05-20 | 2020-05-20 | Face data processing method and device and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111639689B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112885101A (en) * | 2021-03-30 | 2021-06-01 | 浙江大华技术股份有限公司 | Method and device for determining abnormal equipment, storage medium and electronic device |
| CN113255621A (en) * | 2021-07-13 | 2021-08-13 | 浙江大华技术股份有限公司 | Face image filtering method, electronic device and computer-readable storage medium |
| CN115297174A (en) * | 2022-10-10 | 2022-11-04 | 广州市易鸿智能装备有限公司 | Method and device for sequentially storing out-of-order data source, electronic equipment and storage medium |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015036078A (en) * | 2013-08-14 | 2015-02-23 | 株式会社北電子 | Game parlor system, device for game parlor, and method for detecting suspicious person |
| WO2016141641A1 (en) * | 2015-03-11 | 2016-09-15 | 中兴通讯股份有限公司 | Method and device for recognizing suspicious person |
| CN107016322A (en) * | 2016-01-28 | 2017-08-04 | 浙江宇视科技有限公司 | A kind of method and device of trailing personnel analysis |
| WO2017162076A1 (en) * | 2016-03-24 | 2017-09-28 | 北京握奇数据股份有限公司 | Face identification method and system |
| CN107679613A (en) * | 2017-09-30 | 2018-02-09 | 同观科技（深圳）有限公司 | A kind of statistical method of personal information, device, terminal device and storage medium |
| CN107766797A (en) * | 2017-09-27 | 2018-03-06 | 汉柏科技有限公司 | A kind of stranger's recognition methods and system |
| CN108229335A (en) * | 2017-12-12 | 2018-06-29 | 深圳市商汤科技有限公司 | It is associated with face identification method and device, electronic equipment, storage medium, program |
| CN108363947A (en) * | 2017-12-29 | 2018-08-03 | 武汉烽火众智数字技术有限责任公司 | Delay demographic method for early warning based on big data and device |
| CN108509657A (en) * | 2018-04-27 | 2018-09-07 | 深圳爱酷智能科技有限公司 | Data distribute store method, equipment and computer readable storage medium |
| CN109063582A (en) * | 2018-07-10 | 2018-12-21 | 武汉云易虹科技有限公司 | It is a kind of to determine the doubtful system and method bought and sold and divide personnel automatically |
| CN110472491A (en) * | 2019-07-05 | 2019-11-19 | 深圳壹账通智能科技有限公司 | Abnormal face detecting method, abnormality recognition method, device, equipment and medium |
| CN110636258A (en) * | 2019-09-09 | 2019-12-31 | 四川东方网力科技有限公司 | Method, device, equipment and storage medium for analyzing peer personnel |
-
2020
- 2020-05-20 CN CN202010431025.2A patent/CN111639689B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015036078A (en) * | 2013-08-14 | 2015-02-23 | 株式会社北電子 | Game parlor system, device for game parlor, and method for detecting suspicious person |
| WO2016141641A1 (en) * | 2015-03-11 | 2016-09-15 | 中兴通讯股份有限公司 | Method and device for recognizing suspicious person |
| CN107016322A (en) * | 2016-01-28 | 2017-08-04 | 浙江宇视科技有限公司 | A kind of method and device of trailing personnel analysis |
| WO2017162076A1 (en) * | 2016-03-24 | 2017-09-28 | 北京握奇数据股份有限公司 | Face identification method and system |
| CN107766797A (en) * | 2017-09-27 | 2018-03-06 | 汉柏科技有限公司 | A kind of stranger's recognition methods and system |
| CN107679613A (en) * | 2017-09-30 | 2018-02-09 | 同观科技（深圳）有限公司 | A kind of statistical method of personal information, device, terminal device and storage medium |
| CN108229335A (en) * | 2017-12-12 | 2018-06-29 | 深圳市商汤科技有限公司 | It is associated with face identification method and device, electronic equipment, storage medium, program |
| CN108363947A (en) * | 2017-12-29 | 2018-08-03 | 武汉烽火众智数字技术有限责任公司 | Delay demographic method for early warning based on big data and device |
| CN108509657A (en) * | 2018-04-27 | 2018-09-07 | 深圳爱酷智能科技有限公司 | Data distribute store method, equipment and computer readable storage medium |
| CN109063582A (en) * | 2018-07-10 | 2018-12-21 | 武汉云易虹科技有限公司 | It is a kind of to determine the doubtful system and method bought and sold and divide personnel automatically |
| CN110472491A (en) * | 2019-07-05 | 2019-11-19 | 深圳壹账通智能科技有限公司 | Abnormal face detecting method, abnormality recognition method, device, equipment and medium |
| CN110636258A (en) * | 2019-09-09 | 2019-12-31 | 四川东方网力科技有限公司 | Method, device, equipment and storage medium for analyzing peer personnel |
Non-Patent Citations (4)
| Title |
|---|
| FLORIAN SCHROFF等: "FaceNet: A Unified Embedding for Face Recognition and Clustering", PROCEEDINGS OF THE IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION * |
| 史姗姗: "人脸识别技术的研究", 中国优秀硕士学位论文全文数据库 * |
| 汤一平;严海东;: "非约束环境下人脸识别技术的研究", no. 02 * |
| 秦程;: "社区人脸识别的技术途径综述", 重庆工贸职业技术学院学报, no. 02 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112885101A (en) * | 2021-03-30 | 2021-06-01 | 浙江大华技术股份有限公司 | Method and device for determining abnormal equipment, storage medium and electronic device |
| CN112885101B (en) * | 2021-03-30 | 2022-06-14 | 浙江大华技术股份有限公司 | Method and device for determining abnormal equipment, storage medium and electronic device |
| CN113255621A (en) * | 2021-07-13 | 2021-08-13 | 浙江大华技术股份有限公司 | Face image filtering method, electronic device and computer-readable storage medium |
| WO2023284181A1 (en) * | 2021-07-13 | 2023-01-19 | Zhejiang Dahua Technology Co., Ltd. | Method for filtering face images, electronic device, and computer-readable non-transitory storage medium |
| CN115297174A (en) * | 2022-10-10 | 2022-11-04 | 广州市易鸿智能装备有限公司 | Method and device for sequentially storing out-of-order data source, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111639689B (en) | 2023-07-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111639689B (en) | Face data processing method and device and computer readable storage medium | |
| CN108229314B (en) | Target person searching method and device and electronic equipment | |
| Papadopoulos et al. | Cluster-based landmark and event detection for tagged photo collections | |
| US20210357624A1 (en) | Information processing method and device, and storage medium | |
| CN111107319B (en) | Target tracking method, device and system based on regional camera | |
| CN111625687B (en) | Method and system for quickly searching people in media asset video library through human faces | |
| JP2022518459A (en) | Information processing methods and devices, storage media | |
| CN111291682A (en) | Method and device for determining target object, storage medium and electronic device | |
| US20160189395A1 (en) | Information processing apparatus, recording medium, and information processing method | |
| JP2012531130A (en) | Video copy detection technology | |
| CN111209776A (en) | Method, device, processing server, storage medium and system for identifying pedestrians | |
| CN109492604A (en) | Faceform's characteristic statistics analysis system | |
| JP2023527239A (en) | Methods and devices for acquiring information and systems for detecting attention | |
| US20200073877A1 (en) | Video cookies | |
| CN115115825A (en) | Method and device for detecting object in image, computer equipment and storage medium | |
| CN111027385A (en) | Clustering visitor counting method, system, equipment and computer readable storage medium | |
| US11830286B2 (en) | Data processing apparatus, data processing method, and non-transitory storage medium | |
| CN111126112A (en) | Candidate region determination method and device | |
| RU2660599C1 (en) | Method of video data indexing for facet classification | |
| CN112131477A (en) | Library book recommendation system and method based on user portrait | |
| CN115083004B (en) | Identity recognition method and device and computer readable storage medium | |
| CN113454649B (en) | Target detection method, apparatus, electronic device, and computer-readable storage medium | |
| CN114463673B (en) | Material recommendation method, device, equipment and storage medium | |
| Dong et al. | Pavement crack detection based on point cloud data and data fusion | |
| CN110580299A (en) | Method, system, device and storage medium for generating matching of recommendation language of object |
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 |