Nothing Special   »   [go: up one dir, main page]

WO2019062588A1 - 信息识别方法、装置及电子设备 - Google Patents

信息识别方法、装置及电子设备 Download PDF

Info

Publication number
WO2019062588A1
WO2019062588A1 PCT/CN2018/106102 CN2018106102W WO2019062588A1 WO 2019062588 A1 WO2019062588 A1 WO 2019062588A1 CN 2018106102 W CN2018106102 W CN 2018106102W WO 2019062588 A1 WO2019062588 A1 WO 2019062588A1
Authority
WO
WIPO (PCT)
Prior art keywords
target object
historical
current target
detection result
same
Prior art date
Application number
PCT/CN2018/106102
Other languages
English (en)
French (fr)
Inventor
朱碧军
贾海军
李文龙
Original Assignee
阿里巴巴集团控股有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 阿里巴巴集团控股有限公司 filed Critical 阿里巴巴集团控股有限公司
Publication of WO2019062588A1 publication Critical patent/WO2019062588A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/30Scenes; Scene-specific elements in albums, collections or shared content, e.g. social network photos or video
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16Human faces, e.g. facial parts, sketches or expressions
    • G06V40/168Feature extraction; Face representation

Definitions

  • the embodiments of the present invention relate to the field of computer application technologies, and in particular, to an information identification method, device, and electronic device.
  • Face recognition is a kind of biometric recognition technology based on human facial features. Face recognition first needs face detection to determine the face, and then face recognition is performed on the face obtained by the test. For example, the attendance system can detect whether the image includes a face by detecting the collected image; based on the detected face, face recognition is performed from the registered employee database to determine the employee corresponding to the face. Information for the purpose of identity verification.
  • the embodiment of the present invention provides an information identification method, device, and electronic device, which are used to solve the technical problem of low recognition efficiency in the prior art.
  • an information identification method including:
  • it also includes:
  • the current target object that is the same as the historical target object in the history detection result is not recognized.
  • the determining whether the at least one current target object is the same as the historical target object in the historical detection result comprises:
  • the not identifying the current target object that is the same as the historical target object in the historical detection result includes:
  • the determining whether the at least one current target object is the same as the historical target object in the historical detection result comprises:
  • the not identifying the current target object that is the same as the historical target object in the historical detection result includes:
  • determining whether the location area of any current target object is consistent with the location area of any historical target object in the historical detection result includes:
  • the target object is a human face.
  • an information identification apparatus including:
  • a detecting module configured to detect the acquired image to obtain at least one current target object
  • a determining module configured to determine whether the at least one current target object is the same as the historical target object in the historical detection result
  • the first identification module is configured to identify a current target object different from the historical target object in the historical detection result.
  • it also includes:
  • the second identification module is configured to not identify the current target object that is the same as the historical target object in the historical detection result.
  • the determining module is specifically configured to:
  • the second identification module is specifically configured to acquire any current target object and object features of any historical target object in the history detection result that are the same as any of the current target objects; and determine any of the current targets. Whether the object feature of the object is the same as the object feature of any of the same historical target objects; if so, no recognition is performed on any of the current target objects; if not, any of the current target objects are identified.
  • the determining module is specifically configured to: separately acquire an object feature of the at least one current target object; and determine, in the historical detection result, whether the object feature has the same historical target as the object feature of the at least one current target object Object.
  • the second identifying module is specifically configured to determine whether the historical target object that is the same as any current target object is successfully identified in the historical detection result; if yes, the current target object is not identified; No, identify any of the current target objects.
  • the determining module is specifically configured to: determine whether a location offset of a location area of any of the current target objects and a location area of any of the historical target objects in the historical detection result is within a preset range.
  • an embodiment of the present application provides an electronic device, including a processing component, and a memory respectively connected to the processing component;
  • the memory stores one or more computer program instructions, the one or more computer program instructions being invoked and executed by the processing component;
  • the processing component is used to:
  • the processing component detects the acquired image to determine that the at least one current target object is specifically detecting the captured image acquired by the acquisition component to determine at least one current target object.
  • the acquired image is detected to determine at least one current target object; determining whether the at least one current target object is the same as the target object in the historical detection result; only for the target in the historical detection result
  • the current target object with different objects is identified, so the recognition time can be reduced and the recognition efficiency can be improved.
  • FIG. 1 is a flow chart showing an embodiment of an information identification method provided by the present application.
  • FIG. 2 is a flow chart showing still another embodiment of an information identification method provided by the present application.
  • FIG. 3 is a flow chart showing still another embodiment of an information identification method provided by the present application.
  • FIG. 4 is a schematic structural diagram of an embodiment of an information recognition system provided by the present application.
  • FIG. 5 is a schematic structural diagram of still another embodiment of an information recognition system provided by the present application.
  • FIG. 6 is a schematic structural diagram of an embodiment of an information identification apparatus provided by the present application.
  • FIG. 7 is a schematic structural diagram of still another embodiment of an information identifying apparatus provided by the present application.
  • FIG. 8 is a schematic structural diagram of an embodiment of an electronic device provided by the present application.
  • FIG. 9 is a schematic structural diagram of still another embodiment of an electronic device provided by the present application.
  • the technical solution of the embodiment of the present application can be applied to the security field such as attendance, access control, monitoring, security, etc., for identity recognition.
  • the target object in the embodiment of the present application may be a human face, and of course, other biological features that can be human body are not excluded. . Similar to the face recognition process, the recognition of the target object also needs to be first detected to determine the target object, and then the target object obtained by the detection is identified to identify the target object.
  • the technical solution of the embodiment of the present application can reduce the recognition time and improve the recognition efficiency, and is particularly suitable for a scenario in which multiple target objects are simultaneously identified.
  • face recognition in the process of face recognition, image acquisition is always performed, and face detection can be performed for each frame of image, and in a short time, usually hundreds of milliseconds, the object is also collected. That is to say, the position change of the user is usually not large, that is, the images of adjacent frames may be obtained by collecting the same or the same batch of users.
  • face detection and face recognition are performed for each frame of image. However, the detected face may be successfully recognized, thus causing repeated recognition, thereby increasing the recognition time and reducing the recognition efficiency.
  • face recognition is taken as an example. For example, it is necessary to first extract the face feature to construct a face feature template, and then store each face feature template from the database. The comparison is performed to determine the identity information corresponding to the face feature template, and the identity authentication is completed. If the face recognition time can be reduced, the recognition efficiency can be greatly improved.
  • the acquired image is detected to determine at least one current target object; and the at least one current target object is determined whether the target is in the historical detection result.
  • the objects are the same; only the current target object different from the target object in the historical detection result is identified, so the recognition time can be reduced, and the recognition efficiency is improved, especially when there are multiple target objects and need to be recognized at the same time, the recognition efficiency is significantly improved. .
  • FIG. 1 is a flowchart of an embodiment of an information identification method according to an embodiment of the present application, which may include the following steps:
  • the target object detected from the image obtained by the acquisition is named "current target object”.
  • the images obtained for each acquisition can be processed in accordance with the technical solution of the present application.
  • the current detection result is obtained by detecting the acquired image, and the current detection result includes at least one current target object.
  • the target object is a human face
  • the face obtained by the acquisition is subjected to face detection, and the face in the image can be identified by the face detection.
  • the algorithm for extracting the target object from the image may be various, for example, may include a detection algorithm based on histogram rough segmentation and singular value feature, a wavelet transform based detection algorithm, an Adaboost algorithm, etc., and the basic process may be to use the sample image training classification.
  • a detection algorithm based on histogram rough segmentation and singular value feature may include a detection algorithm based on histogram rough segmentation and singular value feature, a wavelet transform based detection algorithm, an Adaboost algorithm, etc.
  • the basic process may be to use the sample image training classification.
  • multiple target objects can be included in an image.
  • the target object in the history detection result is named "historical target object”.
  • Historical test results are obtained by detecting images obtained from historical acquisitions.
  • the historical detection result may specifically refer to the previous detection result, that is, the image obtained by the previous acquisition is detected.
  • the current detection result and the historical detection result are compared, and it is determined whether any current target object is the same as any of the historical target objects in the history detection result.
  • the current target object that is the same as the historical target object in the history detection result is not recognized. That is, in the present embodiment, only the current target object different from the historical target object in the history detection result is identified.
  • the identification is continued for the current target object that is the same as the historical target object in the historical detection result.
  • the identifying the current target object may include, for example, extracting an object feature of the current target object, searching and matching the extracted facial feature with a feature template stored in the database, and calculating a template similarity, if the similarity is greater than the first predetermined value. And determining that the current target object is identified, and the identity information corresponding to the feature template is the identity information of the current target object; if the similarity is less than the second predetermined value, determining that the current target object is not recognized, if similar If the degree is less than the first predetermined value or greater than the second predetermined value, it may be determined that the current target object recognition fails, the identity information of the current target object cannot be identified, and the identification or the like needs to be performed again.
  • the determining whether the at least one current target object is the same as the historical target object in the historical detection result may include:
  • the object feature can be coarse-grained features and extracted by a feature extraction algorithm.
  • the object features are usually represented by multi-dimensional vector data, and the dimension can be lower than the dimension of the object feature extracted by the recognition process.
  • the feature extraction algorithm is, for example, LBP (Local Binary Patterns), a method based on geometric features, a method based on statistical features, and the like, and is the same as the prior art, and details are not described herein again.
  • LBP Local Binary Patterns
  • the object detection can be performed for each frame image, and the detection interval of the adjacent two frames of images is short, usually several tens of milliseconds, the user's position changes even if the user is in a moving state. It is not too small, so it can be judged by the position comparison method whether the target objects in the two detection results are the same target object.
  • FIG. 2 it is a flowchart of still another embodiment of an information identification method provided by an embodiment of the present application, and the method may include the following steps:
  • 201 Detecting the acquired image to obtain at least one current target object.
  • step 202 Determine whether the location area of any current target object is consistent with the location area of any historical target object in the history detection result. If yes, go to step 203, if no, go to step 204.
  • the historical detection result specifically refers to the previous detection result.
  • the acquisition device can automatically search for and capture the image containing the user, and the face recognition image is required, that is, the user's face image is captured.
  • the location area of the target object may be represented by the position coordinate of the preset feature point in the target object, and the target object is a human face, for example, the left eye or the right eye, the nose, and the mouth of the face. Wait.
  • the at least one current target object is directly identified.
  • the not identifying the current target object that is the same as the historical target object in the historical detection result may include:
  • the object feature is further verified in combination with the object feature.
  • the object feature can be a rough feature and is extracted by a feature extraction algorithm.
  • the object feature is usually represented by multi-dimensional vector data, and the dimension can be lower than the dimension of the object feature extracted by the recognition process.
  • the feature extraction algorithm may be, for example, an LBP algorithm, a geometric feature-based method, a statistical feature-based method, or the like, which is the same as the prior art, and is not described herein again.
  • the recognition result obtained by identifying the target object may include the recognition success or the recognition failure, and the recognition success includes the recognition pass or the recognition failure, and the recognition pass, that is, the identity information corresponding to the target object exists in the database, That is, the similarity between the object feature of the target object and the feature template in the database is greater than the first predetermined value; if the identification fails, the identity information corresponding to the target object does not exist in the database, that is, the object feature of the target object and the database The similarity of the feature template is less than the second predetermined value. Failure to identify indicates that the identity information of the target object cannot be confirmed and needs to be re-identified.
  • the current target object also needs to be re-identified to identify the identity information of the current target object, so in order to further ensure the recognition accuracy,
  • the not identifying the current target object that is the same as the historical target object in the historical detection result may include:
  • the recognition success flag may be set for the target object that is successfully identified. Therefore, whether the historical target object that is the same as any current target object in the determination history detection result is successfully identified may be:
  • the current target object that identifies the success may be set to identify the success flag based on the recognition result.
  • the not identifying the current target object that is the same as the historical target object in the historical detection result may include:
  • the current target object whose object number is the same as the object number of the history target object in the history detection result is not recognized.
  • the current target object may not be recognized.
  • the technical solution in the embodiment of the present application can be applied to application fields such as attendance and access control, and is also applicable to network identification in identity identification, security detection and monitoring in important places, identity recognition in smart cards, computer login, and the like. Control and many other different security areas.
  • the target object described in the embodiment of the present application may specifically refer to a human face, and the following uses the target object as a human face as an example to describe the technical solution of the present application.
  • FIG. 3 it is a flowchart of still another embodiment of an information identification method provided by an embodiment of the present application, where the method may include the following steps:
  • step 302 Determine whether the location area of any current face is consistent with the location area of any historical face in the previous detection result. If yes, go to step 303. If no, go to step 304.
  • the location area of the face may refer to a position coordinate of a preset feature point in the face.
  • the preset feature point may be a mouth, a nose, a left eye, or a right eye.
  • the facial feature extraction can be implemented, for example, by using an LBP algorithm.
  • step 306 Determine whether the facial features of any of the current faces are the same as the facial features of any of the historical faces; if yes, perform step 307, and if no, perform step 309.
  • different face numbers can be set for each current face, and a current face with the same historical face as the previous detection result is set to the same face number as any of the historical faces.
  • step 307 Determine whether any of the historical faces is set with a recognition success flag, if yes, execute step 308, and if no, perform step 309.
  • the face recognition process based on the previous detection result, if there is a historical face that is the same as the current face, it may indicate that the current face has been identified, so that it is not necessary to identify again. Reduce recognition time, reduce face recognition time, and improve face recognition efficiency.
  • the information identification system may include an acquisition terminal 401 and an authentication server 402.
  • the collecting terminal 401 is configured to collect an image and send the image to the authentication server 402; the image obtained by the collecting is detected by the authentication server 402 to obtain at least one current target object; and determining whether the at least one current target object is related to the historical detection result.
  • the historical target objects in the same are the same; the current target object that is the same as the historical target object in the history detection result is not recognized; and only the current target object different from the historical target object in the history detection result is identified.
  • image acquisition is implemented by the collection terminal 401, and the object detection and object recognition process are implemented by the authentication server.
  • the collection terminal can perform image collection for a plurality of users located within its collection range, so that a plurality of target objects can be detected from the image, and the target object can be a user's face.
  • the information identification system may include a detection terminal 501 and an authentication server 502;
  • the detecting terminal 501 is configured to collect an image, and detect the acquired image to obtain at least one current target object, and send the at least one current target object to an authentication server; determine whether the at least one current target object is related to history The historical target objects in the detection result are the same; the trigger authentication server 502 does not recognize the current target object that is the same as the historical target object in the history detection result; and triggers the authentication server 502 to be different from the historical target object in the history detection result.
  • the target object is identified.
  • image acquisition and object detection are implemented by the detection terminal, and object recognition is implemented by the authentication server to ensure the processing performance of the detection terminal and the authentication server.
  • the above-mentioned collection terminal, detection terminal or identification terminal can be implemented as an attendance machine with different functions, respectively, to achieve attendance purposes.
  • the attendance time can be recorded corresponding to the identity information.
  • FIG. 6 is a schematic structural diagram of an embodiment of an information identification apparatus according to an embodiment of the present disclosure, where the apparatus may be configured in an authentication server as shown in FIG. 4 or in a detection terminal as shown in FIG. 5. Of course, it can also be configured in the identification terminal.
  • the device can include:
  • the detecting module 601 is configured to detect the acquired image to determine at least one current target object.
  • the determining module 602 is configured to determine whether the at least one current target object is the same as the historical target object in the historical detection result;
  • the first identification module 603 is configured to identify a current target object different from the historical target object in the historical detection result.
  • the device shown in FIG. 6 is different in that the device may further include:
  • the second identification module 604 is configured to identify the current target object that is the same as the historical target object in the historical detection result.
  • the determining module may be specifically configured to: separately acquire an object feature of the at least one current target object; and determine, in the historical detection result, whether the object feature and the at least one current target object are present Historical target objects with the same characteristics.
  • the determining module may be specifically configured to:
  • the determining module may be specifically configured to determine whether a location offset of a location area of any of the current target objects and a location area of any of the historical target objects in the historical detection result is within a preset range.
  • the second identification module may be specifically configured to acquire any current target object and any historical target in the historical detection result that is the same as any of the current target objects.
  • An object feature of the object determining whether the object feature of any one of the current target objects is the same as the object feature of any of the same historical target objects; if so, not identifying any of the current target objects; , identifying any of the current target objects.
  • the object feature is further verified in combination with the object feature.
  • the second identification module may be specifically configured to determine whether the historical target object that is the same as any current target object is successfully identified in the historical detection result; if yes, to any of the current targets The object is not identified; if not, any of the current target objects are identified.
  • the recognition success flag may be set for the target object that is successfully identified, so that the first identification module determines whether the historical target object that is the same as any current target object is successfully identified in the historical detection result. specifically is:
  • the current target object that is successfully identified may be set to identify the success flag based on the recognition result.
  • the first identification module may be specifically configured to set different object numbers for the at least one current target object, where the current detection result is the same target as the historical detection result.
  • the object number of the object is the same;
  • the current target object whose object number is the same as the object number of the history target object in the history detection result is not recognized.
  • the information identifying apparatus described in FIG. 6 or FIG. 7 can perform the information identifying method described in any of the embodiments of FIG. 1 to FIG. 3, and the implementation principle and technical effects thereof are not described again.
  • the specific manner in which the operations are performed by the respective modules and units has been described in detail in the embodiment related to the method, and will not be described in detail herein.
  • the information identifying apparatus of the embodiment shown in FIG. 6 or FIG. 7 can be implemented as an electronic device.
  • the electronic device can include a processing component 801, and the processing component 801, respectively. Connected memory 802;
  • the memory 802 stores one or more computer program instructions that are invoked and executed by the processing component 801;
  • the processing component 801 is configured to:
  • the processing component 801 is further configured to not identify the current target object that is the same as the historical target object in the historical detection result.
  • the processing component does not recognize the current target object that is the same as the historical target object in the history detection result.
  • the electronic device may be an authentication server connected to the collection terminal, and the collection terminal may be an imaging device such as a camera.
  • the embodiment is different from the embodiment shown in FIG. 8 in that the electronic device may further include an acquisition component 803 connected to the processing component 801 for acquiring images.
  • the processing component 801 is specifically configured to detect an image acquired by the acquisition component 803 to determine at least one current target object.
  • the electronic device can be an independent recognition terminal that implements image acquisition, object detection, and object recognition.
  • the current target object determined by the processing component 801 may also be sent to the authentication server, and the processing component 801 does not identify the current target object that is the same as the historical target object in the history detection result, which may be triggered authentication.
  • the server does not recognize the current target object that is the same as the historical target object in the history detection result.
  • the processing component 801 may identify the current target object different from the historical target object in the historical detection result.
  • the trigger authentication server may identify the current target object different from the historical target object in the historical detection result.
  • the processing component 801 can include one or more processors to execute computer instructions to perform all or part of the steps described above.
  • the processing component can also be one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs).
  • ASICs application specific integrated circuits
  • DSPs digital signal processors
  • DSPDs digital signal processing devices
  • PLDs programmable logic devices
  • FPGAs field programmable gate arrays
  • Memory 802 is configured to store various types of data to support operation at the XX device.
  • the memory can be implemented by any type of volatile or non-volatile memory device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), and erasable programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Disk or Optical Disk.
  • SRAM static random access memory
  • EEPROM electrically erasable programmable read only memory
  • EPROM erasable programmable Read Only Memory
  • PROM Programmable Read Only Memory
  • ROM Read Only Memory
  • Magnetic Memory Flash Memory
  • Disk Disk or Optical Disk.
  • Acquisition component 803 can be a camera.
  • the electronic device must also include other components such as an input/output interface, a communication component, and the like.
  • the input/output interface provides an interface between the processing component and the peripheral interface module, and the peripheral interface module may be an output device, an input device, or the like.
  • the communication component is configured to facilitate wired or wireless communication between the electronic device and other devices, such as communicating with an authentication server, and the like.
  • the embodiment of the present application further provides a computer readable storage medium storing a computer program, and when the computer program is executed by a computer, the information identification method of the embodiment shown in any one of the foregoing FIG. 1 to FIG. 3 can be implemented.
  • the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, ie may be located A place, or it can be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment. Those of ordinary skill in the art can understand and implement without deliberate labor.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Image Analysis (AREA)

Abstract

本申请实施例提供一种信息识别方法、装置及电子设备,涉及计算机应用技术领域。其中,对采集获得的图像进行检测,以确定至少一个当前目标对象;判断所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;对与历史检测结果中的历史目标对象不同的当前目标对象进行识别,本申请实施例提供的技术方案减少了不必要的时间,提高了识别效率。

Description

信息识别方法、装置及电子设备
本申请要求2017年09月26日递交的申请号为201710884606.X、发明名称为“信息识别方法、装置及电子设备”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请实施例涉及计算机应用技术领域,尤其涉及一种信息识别方法、装置及电子设备。
背景技术
在现今的考勤、门禁、监控等应用领域中,均涉及快速确认人员身份的需求,而目前通常利用人体的生物特征来进行身份认证,其中,人脸识别应用的最为广泛。
人脸识别是基于人的脸部特征进行身份识别的一种生物识别技术,人脸识别首先需要进行人脸检测,以确定人脸,之后再对检测获得的人脸进行人脸识别,以考勤应用为例,考勤系统通过对采集的图像进行检测,可以确定图像中是否包括人脸;基于检测到的人脸,从已登记的员工数据库中进行人脸识别,以确定该人脸对应的员工信息,实现身份确认的目的。
由于图像采集通常一直进行,针对每一帧图像均会进行人脸识别,而相邻的几帧图像可能均是针对同一个用户采集获得,这就会导致重复工作,而影响识别的效率。
发明内容
本申请实施例提供一种信息识别方法、装置及电子设备,用以解决现有技术中识别效率较低的技术问题。
第一方面,本申请实施例中提供了一种信息识别方法,包括:
对采集获得的图像进行检测,以获得至少一个当前目标对象;
确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;
对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
可选地,还包括:
对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别。
可选地,所述确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对 象相同包括:
确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域是否一致,若是,确定所述任一当前目标对象与所述任一历史目标对象相同,否则,确定所述任一当前目标对象与所述任一历史目标对象不同。
可选地,所述对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别包括:
获取任一当前目标对象以及历史检测结果中与所述任一当前目标对象相同的任一历史目标对象的对象特征;
确定所述任一当前目标对象的对象特征是否和与其相同的所述任一历史目标对象的对象特征相同;
如果是,对所述任一当前目标对象不进行识别;
如果否,对所述任一当前目标对象进行识别。
可选地,所述确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同包括:
分别获取所述至少一个当前目标对象的对象特征;
确定历史检测结果中,是否存在对象特征与所述至少一个当前目标对象的对象特征相同的历史目标对象。
可选地,所述对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别包括:
确定历史检测结果中,与任一当前目标对象相同的历史目标对象是否识别成功;
如果是,对所述任一当前目标对象不进行识别;
如果否,对所述任一当前目标对象进行识别。
可选地,所述确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域是否一致包括:
确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域的位置偏移是否在预设范围。
可选地,所述目标对象为人脸。
第二方面,本申请实施例中提供了一种信息识别装置,包括:
检测模块,用于对采集获得的图像进行检测,以获得至少一个当前目标对象;
判断模块,用于确定所述至少一个当前目标对象是否与历史检测结果中的历史目标 对象相同;
第一识别模块,用于对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
可选地,还包括:
第二识别模块,用于对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别。
可选地,所述判断模块具体用于:
确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域是否一致,若是,确定所述任一当前目标对象与所述任一历史目标对象相同,否则,确定所述任一当前目标对象与所述任一历史目标对象不同。
可选地,所述第二识别模块具体用于获取任一当前目标对象以及历史检测结果中与所述任一当前目标对象相同的任一历史目标对象的对象特征;确定所述任一当前目标对象的对象特征是否和与其相同的所述任一历史目标对象的对象特征相同;如果是,对所述任一当前目标对象不进行识别;如果否,对所述任一当前目标对象进行识别。
可选地,所述判断模块具体用于:分别获取所述至少一个当前目标对象的对象特征;确定历史检测结果中,是否存在对象特征与所述至少一个当前目标对象的对象特征相同的历史目标对象。
可选地,所述第二识别模块具体用于确定历史检测结果中,与任一当前目标对象相同的历史目标对象是否识别成功;如果是,对所述任一当前目标对象不进行识别;如果否,对所述任一当前目标对象进行识别。
可选地,所述判断模块具体用于:确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域的位置偏移是否在预设范围。
第三方面,本申请实施例中提供了一种电子设备,包括处理组件,以及分别与所述处理组件连接的存储器;
所述存储器存储一条或多条计算机程序指令,所述一条或多条计算机程序指令供所述处理组件调用并执行;
所述处理组件用于:
对采集获得的图像进行检测,以确定至少一个当前目标对象;
确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;
对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
可选地,还包括与所述处理组件连接的采集组件,用于采集图像;
所述处理组件对采集获得的图像进行检测,以确定至少一个当前目标对象具体是对所述采集组件采集获得图像进行检测,以确定至少一个当前目标对象。
本申请实施例中,对采集获得的图像进行检测,以确定至少一个当前目标对象;确定所述至少一个当前目标对象是否与历史检测结果中的目标对象相同;仅对与历史检测结果中的目标对象不同的当前目标对象进行识别,因此可以减少识别时间,提高识别效率。
本申请的这些方面或其他方面在以下实施例的描述中会更加简明易懂。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍,显而易见地,下面描述中的附图是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。
图1示出了本申请提供的一种信息识别方法一个实施例的流程图;
图2示出了本申请提供的一种信息识别方法又一个实施例的流程图;
图3示出了本申请提供的一种信息识别方法又一个实施例的流程图;
图4示出了本申请提供的一种信息识别系统一个实施例的结构示意图;
图5示出了本申请提供的一种信息识别系统又一个实施例的结构示意图;
图6示出了本申请提供的一种信息识别装置一个实施例的结构示意图;
图7示出了本申请提供的一种信息识别装置又一个实施例的结构示意图;
图8示出了本申请提供的一种电子设备一个实施例的结构示意图;
图9示出了本申请提供的一种电子设备又一个实施例的结构示意图。
具体实施方式
为了使本技术领域的人员更好地理解本申请方案,下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述。
在本申请的说明书和权利要求书及上述附图中的描述的一些流程中,包含了按照特定顺序出现的多个操作,但是应该清楚了解,这些操作可以不按照其在本文中出现的顺序来执行或并行执行,操作的序号如101、102等,仅仅是用于区分开各个不同的操作, 序号本身不代表任何的执行顺序。另外,这些流程可以包括更多或更少的操作,并且这些操作可以按顺序执行或并行执行。需要说明的是,本文中的“第一”、“第二”等描述,是用于区分不同的消息、设备、模块等,不代表先后顺序,也不限定“第一”和“第二”是不同的类型。
本申请实施例的技术方案可以应用于考勤、门禁、监控、安防等安全领域中,用于身份识别,本申请实施例中的目标对象可以为人脸,当然也不排除可以为人体的其它生物特征。与人脸识别过程类似,针对目标对象的识别同样首先需要进行检测,以确定目标对象,再对检测获得的目标对象进行识别,以对目标对象进行身份确认。
通过本申请实施例的技术方案可以减少识别时间,提高识别效率,特别是适用于针对多个目标对象同时进行识别的场景中。
以人脸识别为例,正如背景技术中所述,在人脸识别过程中,图像采集一直进行,人脸检测可以针对每一帧图像进行,而短时间内,通常数百毫秒,采集对象也即用户的位置变化通常不大,也即相邻几帧图像可能均为采集同一个或者同一批用户而获得的。现有技术中,对每一帧图像均会进行人脸检测以及人脸识别,但是,可能检测出的人脸已经识别成功,因此会造成重复识别,从而增加了识别时间,降低了识别效率。
为了提高识别效率,发明人经过一系列研究发现,由于识别过程非常复杂,以人脸识别为例,例如需要首先提取人脸特征构造人脸特征模板,再从数据库中存储的各个人脸特征模板中进行比对,以确定该人脸特征模板对应的身份信息,完成身份认证,那么如果可以减少人脸识别时间,则可以大大提高识别效率。
据此,提出了本申请的技术方案,在本申请实施例中,对采集获得的图像进行检测,以确定至少一个当前目标对象;确定所述至少一个当前目标对象是否与历史检测结果中的目标对象相同;仅对与历史检测结果中的目标对象不同的当前目标对象进行识别,因此可以减少识别时间,提高识别效率,特别是当存在多个目标对象同时需要进行识别时,将显著提高识别效率。
下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本申请一部分实施例,而不是全部的实施例。基于本申请中的实施例,本领域技术人员在没有作出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
图1是本申请实施例提供的一种信息识别方法一个实施例的流程图,该方法可以包括以下几个步骤:
101:对采集获得的图像进行检测,以获得至少一个当前目标对象。
为了方便描述,将从采集获得的图像检测出的目标对象命名为“当前目标对象”。针对每一次采集获得的图像均可以按照本申请的技术方案进行处理。
对采集获得的图像进行检测,可以获得当前检测结果,该当前检测结果中包括至少一个当前目标对象。
其中,目标对象为人脸时,也即对采集获得的图像进行人脸检测,通过人脸检测可以识别获得图像中的人脸。
从图像中提取目标对象的算法可以有多种,例如可以包括基于直方图粗分割和奇异值特征的检测算法、基于小波变换的检测算法、Adaboost算法等,其基本过程可以是利用样本图像训练分类器,以实现目标对象的检测等,与现有技术相同,在此不再赘述。
在一个实际应用中,图像中可以包括多个目标对象。
102:确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同。
为了方便描述,将历史检测结果中的目标对象命名为“历史目标对象”。
历史检测结果通过对历史采集获得的图像进行检测获得。
其中,该历史检测结果可以具体是指前一次检测结果,也即是对前一次采集获得的图像进行检测获得。
本实施例中,将当前检测结果以及历史检测结果进行比较,判断任一当前目标对象是否与历史检测结果中的任一历史目标对象相同。
其中,判断目标对象是否相同可以有多种实现方式,在下面实施例中会详细进行介绍。
103:将与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
与历史检测结果中的历史目标对象相同的当前目标对象即不进行识别。也即本实施例中仅对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
可选地,对于与历史检测结果中的历史目标对象相同的当前目标对象则继续进行识别。
其中,对当前目标对象进行识别例如可以包括提取当前目标对象的对象特征,将提取获得的人脸特征与数据库中存储的特征模板进行搜索匹配,计算模板相似度,如果相似度大于第一预定值,则可以确定该当前目标对象识别通过,该特征模板对应的身份信息即为该当前目标对象的身份信息;如果相似度小于第二预定值,则可以确定该当前目标对象识别未通过,如果相似度小于第一预定值或者大于第二预定值,则可以确定该当 前目标对象识别失败,无法识别该当前目标对象的身份信息,需要重新进行识别等。
通过本实施例,仅对与历史检测结果中的任一历史目标对象不同的任一当前目标对象进行识别,而如果任一当前目标对象与历史检测结果中的任一历史目标对象相同,表明该当前目标对象已经进行了识别,则可以无需再进行识别,以减少识别时间,提高识别效率。
其中,作为一种可能的实现方式,所述确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同可以包括:
分别获取所述至少一个当前目标对象的对象特征;
确定历史检测结果中,是否存在对象特征与所述至少一个当前目标对象的对象特征相同的历史目标对象。
该对象特征可以为粗粒度特征,采用特征提取算法提取获得,对象特征通常采用多维向量数据表示,其维度可以低于识别过程提取的对象特征的维度。
特征提取算法例如可以LBP(Local Binary Patterns,局部二值模式)、基于几何特征的方法、基于统计特征的方法等,与现有技术相同,在此不再赘述。
作为另一种可能的实现方式,由于对象检测可以针对每一帧图像进行,而对相邻两帧图像的检测间隔很短,通常数十毫秒,因此,即便用户处于移动状态,用户的位置变化也不会不大,因此可以通过位置比对的方式判断前后两次检测结果中的目标对象是否为同一个目标对象。如图2所示,是本申请实施例提供的一种信息识别方法又一个实施例的流程图,该方法可以包括以下几个步骤:
201:对采集获得的图像进行检测,以获得至少一个当前目标对象。
202:确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域是否一致,如果是,执行步骤203,如果否,执行步骤204。
本实施例中,该历史检测结果具体是指前一次检测结果。
由于图像采集一直进行,用户位于采集设备的采集范围内时,采集设备即可以自动搜索并拍摄包含用户的图像,在需要进行人脸识别,即拍摄用户的人脸图像。
可选地,可以是判断任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域的位置偏移是否在预设范围。
其中,目标对象所在位置区域可以以目标对象中预设特征点的位置坐标表示,以目 标对象为人脸为例,该预设特征点例如可以是人脸中的左眼或者右眼、鼻子、嘴巴等。
为了保证识别准确度,还可以判断当前检测时间与历史检测时间的时间差是否在允许范围内,若是,再判断任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域是否一致;否则,则对该至少一个当前目标对象直接进行识别。
203:确定所述任一当前目标对象与历史检测结果中的所述任一历史目标对象相同。
204:确定所述任一当前目标对象与历史检测结果中的所述任一历史目标对象不同。
205:对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别。
206:对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
也即仅将与历史检测结果中的各个目标对象均不同的当前目标对象进行识别。
本实施例中,通过位置比对的方式可以确定前后两次检测结果中的目标对象是否相同,从而针对相同的目标对象可以无需进行识别,则仅对不同的目标对象进行识别,通过减少识别时间,以缩短识别时间,提高识别效率。
其中,为了保证识别准确度,在某些实施例中,所述对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别可以包括:
获取任一当前目标对象以及历史检测结果中与所述任一当前目标对象相同的任一历史目标对象的对象特征;
确定所述任一当前目标对象的对象特征是否与所述任一历史目标对象的对象特征相同;
如果是,对所述任一当前目标对象不进行识别;
如果否,对所述任一当前目标对象进行识别。
也即针对任一当前目标对象以及历史检测结果中与所述任一当前目标对象相同的任一历史目标对象,结合对象特征进一步的进行验证。
该对象特征可以为粗略特征,采用特征提取算法提取获得,对象特征通常采用多维向量数据表示,其维度可以低于识别过程提取的对象特征的维度。
特征提取算法例如可以为LBP算法、基于几何特征的方法、基于统计特征的方法等,与现有技术相同,在此不再赘述。
此外,结合上文描述可知,对目标对象进行识别获得的识别结果可以包括识别成功或者识别失败,识别成功包括识别通过或者识别未通过,识别通过也即数据库中存在目标对象对应的身份信息,也即该目标对象的对象特征与数据库中的特征模板的相似度大 于第一预定值;识别未通过可以认为数据库中不存在目标对象对应的身份信息,也即该目标对象的对象特征与数据库中的特征模板的相似度小于第二预定值。识别失败表明无法确认目标对象的身份信息,需要重新进行识别。
因此,如果历史检测结果中,与当前目标对象相同的历史目标对象识别失败,则该当前目标对象也需要重新进行识别,以识别该当前目标对象的身份信息,因此为了进一步保证识别准确度,在某些实施例中,所述对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别可以包括:
确定历史检测结果中,与任一当前目标对象相同的历史目标对象是否识别成功;
如果是,对所述任一当前目标对象不进行识别;
如果否,对所述任一当前目标对象进行识别。
其中,可选地,为了方便识别,可以对识别成功的目标对象设置识别成功标记,因此所述判断历史检测结果中,与任一当前目标对象相同的历史目标对象是否识别成功可以是:
确定历史检测结果中,与任一当前目标对象相同的历史目标对象是否设置有识别成功标记。
因此,对所述任一当前目标对象进行识别之后,可以基于识别结果,将识别成功的当前目标对象设置识别成功标记。
此外,为了进一步提高识别便利性,在某些实施例中,所述对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别可以包括:
为所述至少一个当前目标对象设置不同对象编号,其中,当前检测结果与历史检测结果中相同的目标对象的对象编号相同;
对对象编号与历史检测结果中的历史目标对象的对象编号相同的当前目标对象不进行识别。
其中,如果历史检测结果中,与任一当前目标对象的对象编号相同的历史目标对象设置有识别成功标记,则即可以对所述任一当前目标对象不进行识别。
本申请实施例中的技术方案,可以应用于考勤、门禁等应用领域中,当然也适用于证件中的身份识别、重要场所中的安全检测和监控、智能卡中的身份识别、计算机登录等网络安全控制等多种不同的安全领域。
其中,在实际应用中,本申请实施例中所述的目标对象可以具体即是指人脸,下面以目标对象为人脸为例,对本申请的技术方案进行描述。
如图3所示,为本申请实施例提供的一种信息识别方法又一个实施例的流程图,该方法可以包括以下几个步骤:
301:对采集获得的图像进行人脸检测,以获得至少一个当前人脸。
302:确定任一当前人脸所在位置区域与前一次检测结果中的任一历史人脸所在位置区域是否一致,如果是执行步骤303,如果否,执行步骤304。
可选地,可以是判断任一当前人脸所在位置区域与前一次检测结果中的任一历史人脸所在位置区域的位置偏移是否在预设范围。
其中,人脸所在位置区域可以是指人脸中某一预设特征点的位置坐标,例如该预设特征点可以为嘴巴、鼻子、左眼或者右眼等。
303:确定所述任一当前人脸与前一次检测结果中的所述任一历史人脸相同,并执行步骤305。
304:确定所述任一当前人脸与前一次检测结果中的所述任一历史人脸不同,并执行步骤309。
305:分别获取所述任一当前人脸以及所述任一历史人脸的脸部特征。
其中,脸部特征提取例如可以采用LBP算法实现。
306:确定所述任一当前人脸的脸部特征是否与所述任一历史人脸的脸部特征相同;如果是,执行步骤307,如果否,执行步骤309。
可选地,可以为各个当前人脸设置不同的人脸编号,保证与前一次检测结果中的任一历史人脸相同的一当前人脸设置与所述任一历史人脸相同的人脸编号。
307:确定所述任一历史人脸是否设置有识别成功标记,如果是,执行步骤308,如果否,执行步骤309。
308:对所述任一当前人脸不进行识别。
309:对所述任一当前人脸进行人脸识别。
310:基于识别结果,将识别成功的各个当前人脸设置识别成功标记。
本实施例中,在人脸识别过程中,基于前一次检测结果,如果存在与当前人脸相同的历史人脸,则可以表明该当前人脸已经经过识别,因此可以无需再次进行识别,从而可以减少识别时间,减少人脸识别时间,提高人脸识别效率。
其中,本申请实施例的技术方案可以应用于信息识别系统中作为一个实施例,如图4中所示,该信息识别系统可以包括采集终端401以及认证服务器402;
采集终端401用于采集图像,并将图像发送至认证服务器402;由认证服务器402对采集获得的图像进行检测,以获得至少一个当前目标对象;确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别;而仅对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
也即由采集终端401实现图像采集,由认证服务器实现对象检测以及对象识别过程。
采集终端可以针对位于其采集范围内的多个用户进行图像采集,从而可以从图像中检测获得多个目标对象,该目标对象可以为用户的人脸。
作为又一个实施例,如图5中所示,该信息识别系统可以包括检测终端501以及认证服务器502;
检测终端501用于采集图像,并对采集获得的图像进行检测,以获得至少一个当前目标对象,并将所述至少一个当前目标对象发送至认证服务器;确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;触发认证服务器502将与历史检测结果中的历史目标对象相同的当前目标对象不进行识别;以及触发认证服务器502将与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
也即由检测终端实现图像采集以及对象检测,由认证服务器实现对象识别,以保证检测终端以及认证服务器的处理性能。
当然,本申请实施例的技术方案也可以应用于独立的识别终端中,由识别终端完成图像采集、对象检测以及对象识别等操作。
在一个实际应用中,上述的采集终端、检测终端或者识别终端可以分别实现为具有不同功能的考勤机,以实现考勤目的。
在考勤应用中,确定目标对象对应身份信息之后,即可以对应该身份信息记录考勤时间等。
图6为本申请实施例提供的一种信息识别装置一个实施例的结构示意图,其中,该装置可以配置在如图4所示的认证服务器中,也可以配置在如图5所示的检测终端中,当然也可以配置在识别终端中。
该装置可以包括:
检测模块601,用于对采集获得的图像进行检测,以确定至少一个当前目标对象。
判断模块602,用于确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;
第一识别模块603,用于将与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
此外,可选地,如图7中所示,与图6所示装置不同之处在于,该装置还可以包括:
第二识别模块604,用于对与历史检测结果中的历史目标对象相同的当前目标对象进行识别。
通过本实施例,仅对与历史检测结果中的任一历史目标对象不同的任一当前目标对象进行识别,而如果任一当前目标对象与历史检测结果中的任一历史目标对象相同,表明该当前目标对象已经进行了识别,则可以无需再进行识别,以减少识别时间,提高识别效率。
作为一种可能的实现方式,所述判断模块可以具体用于:分别获取所述至少一个当前目标对象的对象特征;确定历史检测结果中,是否存在对象特征与所述至少一个当前目标对象的对象特征相同的历史目标对象。
作为另一种可能的实现方式,所述判断模块可以具体用于:
确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域是否一致,若是,确定所述任一当前目标对象与所述任一历史目标对象相同,否则,确定所述任一当前目标对象与所述任一历史目标对象不同。通过位置比对的方式可以确定前后两次检测结果中的目标对象是否相同,从而针对相同的目标对象可以无需进行识别,则仅对不同的目标对象进行识别,通过减少识别时间,以缩短识别时间,提高识别效率。
可选地,所述判断模块可以具体用于确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域的位置偏移是否在预设范围。
其中,为了保证识别准确度,在某些实施例中,所述第二识别模块可以具体用于获取任一当前目标对象以及历史检测结果中与所述任一当前目标对象相同的任一历史目标对象的对象特征;确定所述任一当前目标对象的对象特征是否和与其相同的所述任一历史目标对象的对象特征相同;如果是,对所述任一当前目标对象不进行识别;如果否,对所述任一当前目标对象进行识别。
也即针对任一当前目标对象以及历史检测结果中与所述任一当前目标对象相同的任一历史目标对象,结合对象特征进一步的进行验证。
此外,在某些实施例中,所述第二识别模块可以具体用于确定历史检测结果中,与任一当前目标对象相同的历史目标对象是否识别成功;如果是,对所述任一当前目标对象不进行识别;如果否,对所述任一当前目标对象进行识别。
其中,可选地,为了方便识别,可以对识别成功的目标对象设置识别成功标记,因此所述第一识别模块判断历史检测结果中,与任一当前目标对象相同的历史目标对象是否识别成功可以具体是:
判断历史检测结果中,与任一当前目标对象相同的历史目标对象是否设置有识别成功标记。
因此,所述第二识别模块对所述任一当前目标对象进行识别之后,还可以基于识别结果,将识别成功的当前目标对象设置识别成功标记。
此外,为了进一步提高识别便利性,在某些实施例中,第一识别模块可以具体用于为所述至少一个当前目标对象设置不同对象编号,其中,当前检测结果与历史检测结果中相同的目标对象的对象编号相同;
对对象编号与历史检测结果中的历史目标对象的对象编号相同的当前目标对象不进行识别。
图6或图7所述的信息识别装置可以执行图1~图3任一实施例所述的信息识别方法,其实现原理和技术效果不再赘述。对于上述实施例中的信息识别装置其中各个模块、单元执行操作的具体方式已经在有关该方法的实施例中进行了详细描述,此处将不做详细阐述说明。
在一个可能的设计中,图6或图7所示实施例的信息识别装置可以实现为一电子设备,如图8所示,该电子设备可以包括处理组件801,以及分别与所述处理组件801连接的存储器802;
所述存储器802存储一条或多条计算机程序指令,所述一条或多条计算机程序指令供所述处理组件801调用并执行;
所述处理组件801用于:
对采集获得的图像进行检测,以确定至少一个当前目标对象;
确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;
对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
处理组件801还用于对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别。
处理组件对与历史检测结果中的历史目标对象相同的当前目标对象则不进行识别。
在一个实际应用中,该电子设备可以为与采集终端连接的认证服务器,该采集终端可以为摄像头等摄像设备。
此外,作为又一个实施例,如图9中所示,与图8所示实施例不同之处在于,该电子设备还可以包括与处理组件801连接,用于采集图像的采集组件803。
处理组件801具体是对所述采集组件803采集获得的图像进行检测,以确定至少一个当前目标对象。
该实施例中,该电子设备可以为一个独立的实现图像采集、对象检测以及对象识别的识别终端。
此外,在某些实施例中,该处理组件801确定的当前目标对象还可以发送至认证服务器,处理组件801对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别可以是触发认证服务器将与历史检测结果中的历史目标对象相同的当前目标对象不进行识别。处理组件801对与历史检测结果中的历史目标对象不同的当前目标对象进行识别具体可以是触发认证服务器将与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
其中,处理组件801可以包括一个或多个处理器来执行计算机指令,以完成上述的方法中的全部或部分步骤。当然处理组件也可以为一个或多个应用专用集成电路(ASIC)、数字信号处理器(DSP)、数字信号处理设备(DSPD)、可编程逻辑器件(PLD)、现场可编程门阵列(FPGA)、控制器、微控制器、微处理器或其他电子元件实现,用于执行上述方法。
存储器802被配置为存储各种类型的数据以支持在XX设备的操作。存储器可以由任何类型的易失性或非易失性存储设备或者它们的组合实现,如静态随机存取存储器(SRAM),电可擦除可编程只读存储器(EEPROM),可擦除可编程只读存储器(EPROM),可编程只读存储器(PROM),只读存储器(ROM),磁存储器,快闪存 储器,磁盘或光盘。
采集组件803可以为摄像头。
当然,电子设备必然还可以包括其他部件,例如输入/输出接口、通信组件等。
输入/输出接口为处理组件和外围接口模块之间提供接口,上述外围接口模块可以是输出设备、输入设备等。
通信组件被配置为便于电子设备和其他设备之间有线或无线方式的通信,例如和认证服务器进行通信等。
本申请实施例还提供了一种计算机可读存储介质,存储有计算机程序,所述计算机程序被计算机执行时可以实现上述图1~图3任一项所示实施例的信息识别方法。
所属领域的技术人员可以清楚地了解到,为描述的方便和简洁,上述描述的系统,装置和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
以上所描述的装置实施例仅仅是示意性的,其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的,作为单元显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。本领域普通技术人员在不付出创造性的劳动的情况下,即可以理解并实施。
通过以上的实施方式的描述,本领域的技术人员可以清楚地了解到各实施方式可借助软件加必需的通用硬件平台的方式来实现,当然也可以通过硬件。基于这样的理解,上述技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品可以存储在计算机可读存储介质中,如ROM/RAM、磁碟、光盘等,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行各个实施例或者实施例的某些部分所述的方法。
最后应说明的是:以上实施例仅用以说明本申请的技术方案,而非对其限制;尽管参照前述实施例对本申请进行了详细的说明,本领域的普通技术人员应当理解:其依然可以对前述各实施例所记载的技术方案进行修改,或者对其中部分技术特征进行等同替换;而这些修改或者替换,并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围。

Claims (17)

  1. 一种信息识别方法,其特征在于,包括:
    对采集获得的图像进行检测,以获得至少一个当前目标对象;
    确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;
    对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
  2. 根据权利要求1所述的方法,其特征在于,还包括:
    对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别。
  3. 根据权利要求1或2所述的方法,其特征在于,所述确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同包括:
    确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域是否一致,若是,确定所述任一当前目标对象与所述任一历史目标对象相同,否则,确定所述任一当前目标对象与所述任一历史目标对象不同。
  4. 根据权利要求3所述的方法,其特征在于,所述对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别包括:
    获取任一当前目标对象以及历史检测结果中与所述任一当前目标对象相同的任一历史目标对象的对象特征;
    确定所述任一当前目标对象的对象特征是否和与其相同的所述任一历史目标对象的对象特征相同;
    如果是,对所述任一当前目标对象不进行识别;
    如果否,对所述任一当前目标对象进行识别。
  5. 根据权利要求1或2所述的方法,其特征在于,所述确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同包括:
    分别获取所述至少一个当前目标对象的对象特征;
    确定历史检测结果中,是否存在对象特征与所述至少一个当前目标对象的对象特征相同的历史目标对象。
  6. 根据权利要求2所述的方法,其特征在于,所述对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别包括:
    确定历史检测结果中,与任一当前目标对象相同的历史目标对象是否识别成功;
    如果是,对所述任一当前目标对象不进行识别;
    如果否,对所述任一当前目标对象进行识别。
  7. 根据权利要求3所述的方法,其特征在于,所述确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域是否一致包括:
    确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域的位置偏移是否在预设范围。
  8. 根据权利要求1所述的方法,其特征在于,所述目标对象为人脸。
  9. 一种信息识别装置,其特征在于,包括:
    检测模块,用于对采集获得的图像进行检测,以获得至少一个当前目标对象;
    判断模块,用于确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;
    第一识别模块,用于对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
  10. 根据权利要求9所述的装置,其特征在于,还包括:
    第二识别模块,用于对与历史检测结果中的历史目标对象相同的当前目标对象不进行识别。
  11. 根据权利要求9或10所述的装置,其特征在于,所述判断模块具体用于:
    确定任一当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域是否一致,若是,确定所述任一当前目标对象与所述任一历史目标对象相同,否则,确定所述任一当前目标对象与所述任一历史目标对象不同。
  12. 根据权利要求10所述的装置,其特征在于,所述第二识别模块具体用于获取任一当前目标对象以及历史检测结果中与所述任一当前目标对象相同的任一历史目标对象的对象特征;确定所述任一当前目标对象的对象特征是否和与其相同的所述任一历史目标对象的对象特征相同;如果是,对所述任一当前目标对象不进行识别;如果否,对所述任一当前目标对象进行识别。
  13. 根据权利要求9或10所述的装置,其特征在于,所述判断模块具体用于:分别获取所述至少一个当前目标对象的对象特征;确定历史检测结果中,是否存在对象特征与所述至少一个当前目标对象的对象特征相同的历史目标对象。
  14. 根据权利要求10所述的装置,其特征在于,所述第二识别模块具体用于确定历史检测结果中,与任一当前目标对象相同的历史目标对象是否识别成功;如果是,对所述任一当前目标对象不进行识别;如果否,对所述任一当前目标对象进行识别。
  15. 根据权利要求11所述的装置,其特征在于,所述判断模块具体用于:确定任一 当前目标对象所在位置区域与历史检测结果中的任一历史目标对象所在位置区域的位置偏移是否在预设范围。
  16. 一种电子设备,其特征在于,包括处理组件,以及分别与所述处理组件连接的存储器;
    所述存储器存储一条或多条计算机程序指令,所述一条或多条计算机程序指令供所述处理组件调用并执行;
    所述处理组件用于:
    对采集获得的图像进行检测,以确定至少一个当前目标对象;
    确定所述至少一个当前目标对象是否与历史检测结果中的历史目标对象相同;
    对与历史检测结果中的历史目标对象不同的当前目标对象进行识别。
  17. 根据权利要求16所述的电子设备,其特征在于,还包括与所述处理组件连接的采集组件,用于采集图像;
    所述处理组件对采集获得的图像进行检测,以确定至少一个当前目标对象具体是对所述采集组件采集获得图像进行检测,以确定至少一个当前目标对象。
PCT/CN2018/106102 2017-09-26 2018-09-18 信息识别方法、装置及电子设备 WO2019062588A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710884606.X 2017-09-26
CN201710884606.XA CN109558773B (zh) 2017-09-26 2017-09-26 信息识别方法、装置及电子设备

Publications (1)

Publication Number Publication Date
WO2019062588A1 true WO2019062588A1 (zh) 2019-04-04

Family

ID=65863117

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/106102 WO2019062588A1 (zh) 2017-09-26 2018-09-18 信息识别方法、装置及电子设备

Country Status (3)

Country Link
CN (1) CN109558773B (zh)
TW (1) TW201915830A (zh)
WO (1) WO2019062588A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111582047A (zh) * 2020-04-15 2020-08-25 浙江大华技术股份有限公司 人脸识别验证通行方法及其相关装置

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111275649A (zh) * 2020-02-03 2020-06-12 北京达佳互联信息技术有限公司 图像处理方法、装置、电子设备和存储介质
CN111401463B (zh) * 2020-03-25 2024-04-30 维沃移动通信有限公司 检测结果输出的方法、电子设备及介质
CN111522795A (zh) * 2020-04-23 2020-08-11 北京互金新融科技有限公司 处理数据的方法和装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8121358B2 (en) * 2009-03-06 2012-02-21 Cyberlink Corp. Method of grouping images by face
CN103491351A (zh) * 2013-09-29 2014-01-01 东南大学 违章建筑物的智能视频监控方法
CN104038742A (zh) * 2014-06-06 2014-09-10 上海卓悠网络科技有限公司 一种基于人脸识别技术的门铃系统
CN104581003A (zh) * 2013-10-12 2015-04-29 北京航天长峰科技工业集团有限公司 一种视频回查定位方法
CN106446816A (zh) * 2016-09-14 2017-02-22 北京旷视科技有限公司 人脸识别方法及装置

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2276234A4 (en) * 2008-04-23 2017-11-22 NEC Corporation Image processing device, camera, image processing method, and program
CN102254148A (zh) * 2011-04-18 2011-11-23 周曦 一种多人动态环境下的实时人脸识别方法
CN102880864A (zh) * 2012-04-28 2013-01-16 王浩 一种从流媒体文件中抓拍人脸的方法
US9524418B2 (en) * 2014-06-05 2016-12-20 Promethean Limited Systems and methods for detecting, identifying and tracking objects and events over time
CN105205482B (zh) * 2015-11-03 2018-10-26 北京英梅吉科技有限公司 快速人脸特征识别及姿态估算方法
CN107105310B (zh) * 2017-05-05 2020-07-10 广州盈可视电子科技有限公司 一种视频直播中人物形象替换方法、装置和一种录播系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8121358B2 (en) * 2009-03-06 2012-02-21 Cyberlink Corp. Method of grouping images by face
CN103491351A (zh) * 2013-09-29 2014-01-01 东南大学 违章建筑物的智能视频监控方法
CN104581003A (zh) * 2013-10-12 2015-04-29 北京航天长峰科技工业集团有限公司 一种视频回查定位方法
CN104038742A (zh) * 2014-06-06 2014-09-10 上海卓悠网络科技有限公司 一种基于人脸识别技术的门铃系统
CN106446816A (zh) * 2016-09-14 2017-02-22 北京旷视科技有限公司 人脸识别方法及装置

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111582047A (zh) * 2020-04-15 2020-08-25 浙江大华技术股份有限公司 人脸识别验证通行方法及其相关装置

Also Published As

Publication number Publication date
CN109558773A (zh) 2019-04-02
TW201915830A (zh) 2019-04-16
CN109558773B (zh) 2023-04-07

Similar Documents

Publication Publication Date Title
WO2015165365A1 (zh) 一种人脸识别方法及系统
WO2019033572A1 (zh) 人脸遮挡检测方法、装置及存储介质
TW202006602A (zh) 三維臉部活體檢測方法、臉部認證識別方法及裝置
US9740934B2 (en) Image recognition device and method for registering feature data in image recognition device
KR101322168B1 (ko) 실시간 얼굴 인식 장치
CN110866466B (zh) 一种人脸识别方法、装置、存储介质和服务器
WO2019062588A1 (zh) 信息识别方法、装置及电子设备
US8842880B2 (en) Information processing apparatus, method of controlling information processing apparatus, and storage medium
WO2019042195A1 (zh) 一种人体目标身份识别方法及装置
CN111178252A (zh) 多特征融合的身份识别方法
CN111931548A (zh) 人脸识别系统、建立人脸识别数据的方法及人脸识别方法
JP2009129237A (ja) 画像処理装置及びその方法
Ahmadi et al. Iris recognition system based on canny and LoG edge detection methods
Adhau et al. Iris recognition methods of a blinked eye in nonideal condition
Chaki An efficient two-stage Palmprint recognition using Frangi-filter and 2-component partition method
CN112633250A (zh) 一种人脸识别检测实验方法及装置
Jayachandra et al. Iris Recognition based on Pupil using Canny edge detection and K-Means Algorithm
Chawla et al. A robust algorithm for iris segmentation and normalization using hough transform
Ibrahim Iris recognition using Haar wavelet transform
Shinde et al. Fingerprint Recognition Based on Deep Learning Pre-Train with Our Best CNN Model for Person Identification
US10726259B2 (en) Image processing method and system for iris recognition
Sehgal Palm recognition using LBP and SVM
Aydoğdu et al. A study on liveness analysis for palmprint recognition system
Hernandez-Garcia et al. Iris recognition using supervised learning based on matching features
Tinuke Omolewa et al. Framework for User Authentication at a Distance for Mobile Phones Using Contactless Hand-based Multimodal Biometric System.

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18860679

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18860679

Country of ref document: EP

Kind code of ref document: A1