WO2013082957A1 - 人脸识别方法及能识别人脸的移动终端 - Google Patents

人脸识别方法及能识别人脸的移动终端 Download PDF

Info

Publication number
WO2013082957A1
WO2013082957A1 PCT/CN2012/081281 CN2012081281W WO2013082957A1 WO 2013082957 A1 WO2013082957 A1 WO 2013082957A1 CN 2012081281 W CN2012081281 W CN 2012081281W WO 2013082957 A1 WO2013082957 A1 WO 2013082957A1
Authority
WO
WIPO (PCT)
Prior art keywords
facial feature
interval
matching
face
target person
Prior art date
Application number
PCT/CN2012/081281
Other languages
English (en)
French (fr)
Inventor
王亚辉
Original Assignee
惠州Tcl移动通信有限公司
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 惠州Tcl移动通信有限公司 filed Critical 惠州Tcl移动通信有限公司
Publication of WO2013082957A1 publication Critical patent/WO2013082957A1/zh

Links

Images

Classifications

    • 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
    • G06V40/172Classification, e.g. identification

Definitions

  • the present invention relates to the field of face recognition, and in particular to a face recognition method and a mobile terminal capable of recognizing a face.
  • the prior art provides a face recognition method, which uses a camera to adopt an image containing a face in a field of view, and then sends the collected image to a computer for image processing to obtain a face. The feature value is finally compared with the feature value in the database to identify the face.
  • the camera collects images at one time. If the target is far away from the camera, the image is blurred, or the light is too dark and the image is overexposed or the brightness is not enough, resulting in blurred images. Clearly, even if the angle of deflection of the target person and the camera exceeds a defined angle, the feature value extraction of the face is insufficient, which will cause the recognition to fail, and the target person may be lucky to escape the monitoring.
  • the camera's field of view can only cover a certain range, and the target can easily evade monitoring if it deliberately bypasses the camera's field of view.
  • the technical problem to be solved by the present invention is to provide a face recognition method and a mobile terminal capable of recognizing a face, which can effectively improve the success rate of face recognition.
  • a technical solution adopted by the present invention is to provide a mobile terminal capable of recognizing a human face, wherein the mobile terminal includes: a camera for collecting an image including a human face; and an identification module, a facial feature point for acquiring a face in the image, if the number of feature points is less than a set threshold, outputting a facial feature invalid code; if the number of feature points is greater than or equal to a set threshold, outputting a facial feature code; And matching the output facial feature code in the facial feature database, and determining whether the similarity between the output facial feature code and the facial feature code in the facial feature code database belongs to a matching interval, a non-matching interval, and a to-be-identified interval.
  • the processor sends a prompt message identifying the target person, and if it is determined to belong to the unmatched interval, the processor issues a prompt message of the non-target person, and if it is determined to belong to the to-be-identified interval, the processor controls the camera to re-acquire the content.
  • the image inside the face until the processor detects the end command or sends out the identified target person.
  • the driving LED light turns red; when the matching result belongs to the matching interval or the unmatched interval, the driving LED light turns green; the matching result belongs to When the interval is recognized, the drive LED turns blue; when the end command is detected, the drive LED turns off.
  • the mobile terminal further includes: a radio frequency module, configured to send a request to update a database command to the server, and receive data transmitted by the server when receiving an update database command sent by the server.
  • a radio frequency module configured to send a request to update a database command to the server, and receive data transmitted by the server when receiving an update database command sent by the server.
  • the radio frequency module is further configured to transmit facial feature code data to the server to update facial feature code data in the server.
  • a face recognition method which includes: collecting an image including a human face; acquiring facial features of a face in the image; Matching the acquired facial features in the feature database, determining which one of the matching interval, the unmatched interval, and the to-be-identified interval is the result of the matching; if the determination belongs to the matching interval, issuing the prompt information identifying the target person; if the determination is a mismatch The interval, the prompt information of the non-target person is sent; if it is determined to belong to the to-be-identified interval, the step of performing the process of collecting the image including the face is performed, until the end instruction is detected or the prompt information identifying the target person or the prompt of the non-target person is issued. information.
  • the step of acquiring a facial feature of a face in an image includes: acquiring a facial feature point of a face in the image; and after the step of matching the acquired facial feature in the facial feature database, the method includes: if the number of feature points is less than a threshold value is output, and a facial feature invalidation code is output; if the number of feature points is greater than or equal to the set threshold, the facial feature code is output; and the step of matching the acquired facial feature in the facial feature database includes: Matching the output facial feature code in the database; the step of determining the matching result as belonging to the matching interval, the unmatched interval, and the to-be-identified interval includes: determining the output facial feature code and the face in the facial feature code database The similarity of the feature codes belongs to which of the matching interval, the unmatched interval, and the to-be-identified interval.
  • the step of matching the output facial feature code in the facial feature database further includes the following steps: sending a request to update the database command to the server; receiving the update database command sent by the server, receiving the facial feature code data transmitted by the server, and The facial signature data is downloaded to update the facial signature database.
  • the step of outputting the facial feature code further comprises the step of transmitting facial feature code data to the server to update facial feature code data in the server.
  • a mobile terminal capable of recognizing a human face
  • the mobile terminal includes: a camera for collecting an image including a human face; and an identification module a method for acquiring a facial feature of a face in the image; a processor, configured to match the acquired facial feature in the facial feature database, and determine which of the matching interval, the unmatched interval, and the to-be-identified segment If the judgment belongs to the matching interval, the processor sends a prompt message identifying the target person, and if it is determined to belong to the unmatched interval, the processor issues a prompt message of the non-target person, and if it is determined to belong to the to-be-identified interval, the processor controls the camera to re-collect the included person. The image of the face until the processor detects the end instruction or issues a prompt message identifying the target person or a prompt message of the non-target person.
  • the identification module is further configured to acquire facial feature points of the face in the image, and if the number of the feature points is less than the set threshold, output the facial feature invalid code; if the number of the feature points is greater than or equal to the set threshold And outputting a facial feature code; the processor is further configured to match the output facial feature code in the facial feature database, and determine that the similarity between the output facial feature code and the facial feature code in the facial feature code database is matched Which of the interval, the unmatched interval, and the interval to be identified.
  • the mobile terminal further includes: a radio frequency module, configured to send a request to update a database command to the server, and receive data transmitted by the server when receiving an update database command sent by the server.
  • a radio frequency module configured to send a request to update a database command to the server, and receive data transmitted by the server when receiving an update database command sent by the server.
  • the radio frequency module is further configured to transmit facial feature code data to the server to update facial feature code data in the server.
  • the processor is further configured to drive the LED light to turn red when the face recognition is initiated; when the matching result belongs to the matching interval or the unmatched interval, the driving LED light turns green; when the matching result belongs to the to-be-identified interval The drive LED light turns blue; when the end command is detected, the drive LED light goes out.
  • the beneficial effects of the present invention are: different from the prior art, when the image captured by the camera is blurred or the angle of the target person and the camera is deflected exceeds a defined angle, the result of the matching is located in the interval to be identified, that is, the image cannot be When the recognition is not a perfect match or a complete mismatch, the processor will reacquire the image until the image is accurately recognized, improving the recognition rate.
  • the present invention combines the face recognition technology and the mobile terminal technology to collect images of the target person at any time and any place through the mobile terminal, thereby effectively expanding the field of view of the camera.
  • FIG. 1 is a flow chart of a first embodiment of a face recognition method according to the present invention.
  • FIG. 2 is a schematic diagram of distribution of feature points in a face of a first embodiment of a face recognition method according to the present invention
  • FIG. 3 is a schematic structural view of a first embodiment of a face recognition device according to the present invention.
  • FIG. 4 is a schematic structural diagram of Embodiment 1 of a mobile terminal capable of recognizing a human face according to the present invention.
  • a first embodiment of a face recognition method according to the present invention includes:
  • Step 110 Acquire an image including a human face.
  • Step 120 Acquire a facial feature of a face in the image.
  • facial features include: black sputum, birthmark, distance from the tip of the nose to the eyes, height of the bridge of the nose, length and width, distance of the tibia, eyes, jaw contour, brow bone and other features with obvious recognition.
  • Step 130 Match the acquired facial features in the facial feature database, and determine which of the matching interval, the unmatched interval, and the to-be-identified segment are matched.
  • the matching interval, the to-be-identified interval, and the unmatched interval may be set to be continuous or discontinuous at the boundary of the interval, and the specific interval setting is subject to actual requirements.
  • Step 140 If it is determined that it belongs to the matching interval, the prompt information identifying the target person is sent; if it is determined to belong to the unmatched interval, the prompt information of the non-target person is sent; if it is determined to belong to the to-be-identified interval, returning to the image for collecting the face is performed. The steps until the end instruction is detected or the prompt information identifying the target person or the prompt information of the non-target person is issued.
  • the prompt information may be text of a screen, a pattern display, or information displayed by sound and light.
  • the present invention combines the face recognition technology and the mobile terminal technology to collect images of the target person at any time and any place through the mobile terminal, thereby effectively expanding the field of view of the camera.
  • a more detailed face recognition method is illustrated by using a mobile phone as an example but not limited to a mobile phone.
  • a mobile terminal having the camera can be used as the image capturing tool of the present invention, in order to make the recognition rate of the recognized image, a higher pixel camera can be employed.
  • the following uses a mobile phone as an example to describe:
  • the user collects an image of the face of the suspicious person through the mobile phone, and processes the image by histogram equalization and image smoothing to obtain a clear image. Then, the image is normalized to obtain a standard size image. Since histogram equalization, image smoothing, and normalization are all prior art, no further description is made here.
  • an image of a face containing a suspicious person is subjected to image processing techniques such as binarization, edge extraction, and singularity calculation, and facial features of the suspicious person, including black sputum, birthmark, nose tip and eyes.
  • image processing techniques such as binarization, edge extraction, and singularity calculation, and facial features of the suspicious person, including black sputum, birthmark, nose tip and eyes.
  • the distance, the height of the bridge of the nose, the length and width, the distance between the humerus and the eyes, the contour of the jaw, the brow bone and other features with obvious recognition are transformed into feature points.
  • the threshold is the number of feature points with the smallest amount of calculation obtained under the condition that the recognition rate is guaranteed by the computer.
  • the facial feature database in the mobile phone stores a plurality of facial feature codes of the target person, reads the facial feature code of the target person, and matches the output facial feature code with the facial feature code of the target person in the facial feature database.
  • the matching result can be reflected as the similarity.
  • the similarity interval of [0%, 100%] is divided into three intervals according to the similarity degree, the matching interval, the to-be-identified interval and the unmatched interval.
  • the matching interval, the to-be-identified interval, and the unmatched interval may be set to be continuous or discontinuous in the boundary of the interval, for example, [0%, 30%] is divided into unmatched intervals, and [30%, 70%] is divided into to-be-identified.
  • Interval [70%, 100%] is divided into matching intervals, or [0%, 25%] is divided into unmatched intervals, [30%, 70%] is divided into intervals to be identified, [75%, 100%]
  • the specific interval setting is subject to the actual requirements.
  • the suspicious person is the target person, and the processor sends a prompt message identifying the target person, and prompts the user to match through the text, the pattern display or the sound and light alarm on the screen; if the matching is similar The degree belongs to the mismatch interval, the processor issues the prompt information of the non-target person, and reminds the user that the matching is successful through the text, the pattern display or the sound and light alarm on the screen, and the user can abandon the tracking of the suspicious person; if the matching obtains the similarity belongs to The identification interval indicates that the identity of the suspicious person cannot be determined, and the mobile phone prompts through the user interface whether the image of the suspicious person needs to be further collected.
  • the program ends. Otherwise, continue to collect images of the suspicious person until the mobile phone sends a prompt message identifying the target person or a non-target person's prompt information.
  • the facial feature code acquired by the mobile phone may be stored in the facial feature database for expanding the amount of data in the facial feature database, or may be connected to the server through the base station, and the acquired facial feature code is transmitted to the server through the wireless network, and updated. Facial signature data in the server.
  • the server may send a request to update the database instruction, and after receiving the update database command sent by the server, The facial feature code data transmitted by the server is received, and the facial feature code data is downloaded to update the facial feature code database.
  • the present invention combines the face recognition technology and the mobile terminal technology to collect images of the target person at any time and any place through the mobile terminal, thereby effectively expanding the field of view of the camera.
  • the present invention also provides a face recognition device, including:
  • the collecting module 110 is configured to collect an image including a human face
  • the obtaining module 120 is configured to acquire a facial feature of a face in the image
  • the determining module 130 is configured to match the acquired facial features in the facial feature database, and determine which one of the matching interval, the unmatched interval, and the to-be-identified interval is the result of the matching;
  • the prompting module 140 is configured to: when determining that the matching section belongs to the prompting area, issue prompt information for identifying the target person; and when determining that the non-targeting section belongs to, the prompting information of the non-target person is issued.
  • the face recognition device of the present invention can collect an image including a human face of a suspicious person through the acquisition module 110, and transmit the image to the acquisition module 120 to extract facial features of the suspicious person, such as black eyes, birthmarks, and nose tips.
  • facial features of the suspicious person such as black eyes, birthmarks, and nose tips.
  • the determining module 130 reads the facial features of the target person from the facial feature database, and matches the facial features output by the obtaining module 120 with the facial features of the target person in the facial feature database.
  • the matching result can be reflected as the similarity.
  • the similarity interval of [0%, 100%] is divided into three intervals according to the similarity degree, the matching interval, the to-be-identified interval and the unmatched interval.
  • the matching interval, the to-be-identified interval, and the unmatched interval may be set to be continuous or discontinuous in the boundary of the interval, for example, [0%, 30%] is divided into unmatched intervals, and [30%, 70%] is divided into to-be-identified.
  • Interval [70%, 100%] is divided into matching intervals, or [0%, 25%] is divided into unmatched intervals, [30%, 70%] is divided into intervals to be identified, [75%, 100%] To match the interval.
  • the specific interval setting is subject to the actual requirements.
  • the suspicious person is the target person
  • the prompting module 140 issues the prompt information for identifying the target person, and prompts the user to match through the text, the pattern display or the sound and light alarm of the screen; if the matching is obtained
  • the similarity belongs to the mismatch interval, and the prompting module 140 issues the prompt information of the non-target person, and reminds the user that the matching is successful through the text, the pattern display or the sound and light alarm of the screen, and the user can give up the tracking of the suspicious person; if the matching obtains the similarity It belongs to the interval to be identified, indicating that the identity of the suspicious person cannot be determined, and the user interface is prompted whether the image of the suspicious person needs to be further collected.
  • the face recognition device stops working. Otherwise, the collection module 110 continues to collect images of the faces containing the suspicious characters until the prompting module 140 issues prompt information identifying the target person or prompt information of the non-target person.
  • the present invention combines the face recognition technology and the mobile terminal technology to collect images of the target person at any time and any place through the mobile terminal, thereby effectively expanding the field of view of the camera.
  • the present invention further provides a mobile terminal capable of recognizing a human face.
  • the following is still exemplified by a mobile phone, but is not limited to a mobile phone.
  • the pixel of the camera 210 of the mobile phone can select a higher pixel to ensure image recognition. Requirements:
  • the face recognition program When a suspicious person is found, the face recognition program is activated, and the processor 230 drives the LED to turn red, prompting the user that the face recognition program is running. At this time, the user can collect an image of the face containing the suspicious person through the camera 210 of the mobile phone, and transmit the image to the recognition module 220.
  • the recognition module 220 processes the image by histogram equalization, image smoothing, etc. to obtain a clear image. Then, the image is normalized to obtain a standard size image. Since histogram equalization, image smoothing, and normalization are all prior art, no further description is made here.
  • the recognition module 220 images the face of the face containing the suspicious person through facial image processing techniques such as binarization, edge extraction, and calculation of singularities, such as black sputum, birthmark, nose tip to both eyes. The distance, nose height, length, width, humerus, distance between the eyes, mandibular contour, brow bone and other features with obvious recognition are transformed into feature points.
  • the recognition module 220 If the number of acquired feature points is less than a set threshold due to image blur or face deflection, etc., the recognition module 220 outputs a face feature invalidation code. On the contrary, if the acquired number of feature points is greater than or equal to the set threshold, the identification module 220 outputs the facial feature code.
  • the threshold is the number of feature points with the smallest amount of calculation obtained under the condition that the recognition rate is guaranteed by the computer.
  • the processor 230 After receiving the facial feature code output by the identification module 220, the processor 230 reads the facial feature code of the target person from the facial feature database, and performs the facial feature output by the recognition module 220 and the facial feature of the target person in the facial feature database. One by one match.
  • the matching result can be reflected as the similarity.
  • the similarity interval of [0%, 100%] is divided into three intervals according to the similarity degree, the matching interval, the to-be-identified interval and the unmatched interval.
  • the matching interval, the to-be-identified interval, and the unmatched interval may be set to be continuous or discontinuous in the boundary of the interval, for example, [0%, 30%] is divided into unmatched intervals, and [30%, 70%] is divided into to-be-identified.
  • Interval [70%, 100%] is divided into matching intervals, or [0%, 25%] is divided into unmatched intervals, [30%, 70%] is divided into intervals to be identified, [75%, 100%] To match the interval.
  • the specific interval setting is subject to the actual requirements.
  • the processor 230 sends the prompt information identifying the target person, and drives the LED light to turn green to remind the user that the matching is successful; if the similarity obtained by the matching belongs to the mismatch In the interval, the processor 230 sends a prompt message of the non-target person, and drives the LED light to turn green to remind the user that the matching is successful, and the user can abandon the tracking of the suspicious person; if the similarity obtained by the matching belongs to the to-be-identified interval, the suspicious person cannot be determined.
  • the identity of the processor 230 driving the LED light turns blue to remind the user to further collect images of the suspicious person, and the mobile phone prompts through the user interface whether further images of the suspicious person need to be collected. If the mobile phone detects that the user has sent an end command, or because the timeout mobile phone system automatically issues an end command, ending the program, the processor 230 drives the LED to go out. Otherwise, continue to collect images of the suspicious person until the mobile phone sends a prompt message identifying the target person or a non-target person's prompt information.
  • the facial feature code acquired by the mobile phone may be stored in the facial feature database of the local device for expanding the amount of data in the facial feature database, or may be connected to the server through the radio frequency module 240, and the acquired facial feature code is transmitted through the wireless network. Go to the server and update the facial signature data in the server.
  • the RF module 240 sends a request to update the database instruction to the server, and receives the update sent by the server. After the database command, the facial feature code data transmitted by the server is received, and the facial feature code data is downloaded to update the facial feature code database.
  • the radio frequency module 240 can be one or more of GSM, GPRS, TDS-CDMA, CDMA2000, WCDMA.
  • the processor 230 will reacquire the image until the image is accurately recognized.
  • the present invention combines the face recognition technology and the mobile terminal technology to collect images of the target person at any time and any place through the mobile terminal, thereby effectively expanding the field of view of the camera.

Abstract

本发明公开了人脸识别方法及装置以及能识别人脸的移动终端。其中,所述方法包括:采集包含人脸在内的图像;获取图像中人脸的面部特征;在面部特征数据库中匹配获取的面部特征,判断匹配的结果属于匹配区间、不匹配区间及待识别区间的哪一个;若判断属于匹配区间,发出识别出目标人物的提示信息;若判断属于不匹配区间,发出非目标人物的提示信息;若判断属于待识别区间,返回执行采集包含人脸在内的图像的步骤,直到检测到结束指令或发出识别出目标人物的提示信息或非目标人物的提示信息。

Description

人脸识别方法及能识别人脸的移动终端
【技术领域】
本发明涉及人脸识别领域,特别是涉及一种人脸识别方法及能识别人脸的移动终端。
【背景技术】
众所周知,除了双胞胎及多胞胎外,出现两张一模一样的脸的概率还不到七十万亿分之一,即便是双胞胎及多胞胎也会受后天环境、气候、生活的影响而出现外貌的差异,因而,人脸识别技术是生物特征识别技术的一个主要方向。
为了方便的对人脸进行识别,现有技术提供了一种人脸识别方法,利用摄像头采视场中包含人脸的图像,然后将采集到的图像送入计算机中进行图像处理获得人脸的特征值,最后与数据库中的特征值进行比较,进而识别人脸。
但是,在该方案中,摄像头为一次性采集图像,如果采集时,由于目标人物离摄像比较远而导致图像模糊不清,或者光线过明过暗使得图像过度曝光或者亮度不够而导致图像模糊不清,甚至,目标人物与摄像头偏转的角度超过限定的角度造成人脸的特征值提取不足够,将会导致该次识别失败,目标人物就有可能侥幸地逃避监测。
此外,摄像头的视场只能覆盖一定的范围,如果目标人物刻意绕过摄像头的视场则可轻易地逃避监测。
【发明内容】
本发明主要解决的技术问题是提供一种人脸识别方法及能识别人脸的移动终端,能够有效提高人脸识别的成功率。
为解决上述技术问题,本发明采用的一个技术方案是:提供一种能识别人脸的移动终端,其中,所述移动终端包括:摄像头,用于采集包含人脸在内的图像;识别模块,用于获取图像中人脸的面部特征点,若特征点的数量小于设定的阈值,则输出面部特征无效码;若特征点的数量大于或等于设定的阈值,则输出面部特征码;处理器,用于在面部特征数据库中匹配输出的面部特征码,判断所述输出的面部特征码与面部特征码数据库中的面部特征码的相似度属于匹配区间、不匹配区间及待识别区间的哪一个;若判断属于匹配区间,处理器发出识别出目标人物的提示信息,若判断属于不匹配区间,处理器发出非目标人物的提示信息,若判断属于待识别区间,处理器控制摄像头重新采集包含人脸在内的图像,直到处理器检测到结束指令或发出识别出目标人物的提示信息或非目标人物的提示信息,以及在启动人脸识别时,驱动LED灯变成红色;在匹配结果属于匹配区间或不匹配区间时,驱动LED灯变成绿色;在匹配结果属于待识别区间时,驱动LED灯变成蓝色;在检测到结束命令时,驱动LED灯熄灭。
其中,所述移动终端还包括:射频模块,用于向服务器发送请求更新数据库指令及在接收到服务器发送的更新数据库指令时,接收服务器传输的数据。
其中,所述射频模块还用于向所述服务器传输面部特征码数据以更新所述服务器中的面部特征码数据。
为解决上述技术问题,本发明采用的另一个技术方案是:提供一种人脸识别方法,其中,包括:采集包含人脸在内的图像;获取所述图像中人脸的面部特征;在面部特征数据库中匹配所述获取的面部特征,判断匹配的结果属于匹配区间、不匹配区间及待识别区间的哪一个;若判断属于匹配区间,发出识别出目标人物的提示信息;若判断属于不匹配区间,发出非目标人物的提示信息;若判断属于待识别区间,返回执行采集包含人脸在内的图像的步骤,直到检测到结束指令或发出识别出目标人物的提示信息或非目标人物的提示信息。
其中,所述获取图像中人脸的面部特征的步骤包括:获取图像中人脸的面部特征点;在所述面部特征数据库中匹配获取的面部特征的步骤之后,包括:若特征点的数量小于设定的阈值,则输出面部特征无效码;若特征点的数量大于或等于设定的阈值,则输出面部特征码;所述在面部特征数据库中匹配获取的面部特征的步骤包括:在面部特征数据库中匹配输出的面部特征码;所述判断匹配的结果为属于匹配区间、不匹配区间及待识别区间的哪一个的步骤包括:判断所述输出的面部特征码与面部特征码数据库中的面部特征码的相似度属于匹配区间、不匹配区间及待识别区间的哪一个。
其中,所述在面部特征数据库中匹配输出的面部特征码步骤之前还包括如下步骤:向服务器发送请求更新数据库指令;若接收到服务器发送的更新数据库指令,接收服务器传输的面部特征码数据,并下载所述面部特征码数据以更新所述面部特征码数据库。
其中,所述在特征点的数量大于或等于设定的阈值时,输出面部特征码步骤之后还包括如下步骤:向所述服务器传输面部特征码数据以更新所述服务器中的面部特征码数据。
为解决上述技术问题,本发明采用的又一个技术方案是:提供一种能识别人脸的移动终端,其中,所述移动终端包括:摄像头,用于采集包含人脸在内的图像;识别模块,用于获取所述图像中人脸的面部特征;处理器,用于在面部特征数据库中匹配所述获取的面部特征,判断匹配的结果属于匹配区间、不匹配区间及待识别区间的哪一个;若判断属于匹配区间,处理器发出识别出目标人物的提示信息,若判断属于不匹配区间,处理器发出非目标人物的提示信息,若判断属于待识别区间,处理器控制摄像头重新采集包含人脸在内的图像,直到处理器检测到结束指令或发出识别出目标人物的提示信息或非目标人物的提示信息。
其中,所述识别模块还进一步用于获取图像中人脸的面部特征点,若特征点的数量小于设定的阈值,则输出面部特征无效码;若特征点的数量大于或等于设定的阈值,则输出面部特征码;所述处理器还进一步用于在面部特征数据库中匹配输出的面部特征码,判断所述输出的面部特征码与面部特征码数据库中的面部特征码的相似度属于匹配区间、不匹配区间及待识别区间的哪一个。
其中,所述移动终端还包括:射频模块,用于向服务器发送请求更新数据库指令及在接收到服务器发送的更新数据库指令时,接收服务器传输的数据。
其中,所述射频模块还用于向所述服务器传输面部特征码数据以更新所述服务器中的面部特征码数据。
其中,所述处理器还用于在启动人脸识别时,驱动LED灯变成红色;在匹配结果属于匹配区间或不匹配区间时,驱动LED灯变成绿色;在匹配结果属于待识别区间时,驱动LED灯变成蓝色;在检测到结束命令时,驱动LED灯熄灭。
本发明的有益效果是:区别于现有技术的情况,当摄像头拍摄到的图像模糊不清或目标人物与摄像头偏转的角度超过限定的角度而造成匹配的结果位于待识别区间,即图像不能被识别但又不属于完全匹配或者完全不匹配的情况时,处理器会重新采集图像,直到图像能够准确被识别,提高识别率。并且,本发明结合人脸识别技术及移动终端技术,通过移动终端随时随地对目标人物进行采集图像,有效的扩大摄像头的视场。
【附图说明】
图1是本发明一种人脸识别方法实施例一的流程图;
图2是本发明一种人脸识别方法实施例一的人脸中特征点分布示意图;
图3是本发明一种人脸识别装置实施例一的结构示意图;
图4是本发明一种能够识别人脸的移动终端实施例一的结构示意图。
【具体实施方式】
下面结合附图和实施例对本发明进行详细说明。
参阅图1,本发明一种人脸识别方法实施例一包括:
步骤110:采集包含人脸在内的图像。
步骤120:获取图像中人脸的面部特征。其中,面部特征包括:黑痣、胎记、鼻尖到双眼的距离、鼻梁的高度、长度及宽度、颧骨、双眼的距离、下颌轮廓、眉骨及其它具有明显识别作用的特征。
步骤130:在面部特征数据库中匹配获取的面部特征,判断匹配的结果属于匹配区间、不匹配区间及待识别区间的哪一个。其中,匹配区间、待识别区间及不匹配区间可以设置在区间的边界连续或者不连续,具体的区间设置以实际要求为准。
步骤140:若判断属于匹配区间,发出识别出目标人物的提示信息;若判断属于不匹配区间,发出非目标人物的提示信息;若判断属于待识别区间,返回执行采集包含人脸在内的图像的步骤,直到检测到结束指令或发出识别出目标人物的提示信息或非目标人物的提示信息。提示信息可以是屏幕的文字、图案显示或者通过声光显示的信息。
区别于现有技术的情况,当摄像头拍摄到的图像模糊不清或目标人物与摄像头偏转的角度超过限定的角度而造成匹配的结果位于待识别区间,即图像不能被识别但又不属于完全匹配或者完全不匹配的情况时,处理器会重新采集图像,直到图像能够准确被识别,提高识别率。并且,本发明结合人脸识别技术及移动终端技术,通过移动终端随时随地对目标人物进行采集图像,有效的扩大摄像头的视场。
下面,以手机为例但不限于手机阐述更详尽的人脸识别方法。只要具有摄像头的移动终端都可以作为本发明的图像采集工具,为了使识别的图像的识别率,可以采用较高像素的摄像头。为了描述方便,下面以手机为例来描述:
发现可疑人物时,用户通过手机采集一幅可疑人物的人脸的图像,对图像采用直方图均衡化、图像平滑等方法进行处理得到清晰的图像。然后,对图像进行归一化处理,获得标准尺寸的图像。由于直方图均衡化、图像平滑及归一化均为现有技术,这里就不展开赘述。
参阅图2,在本实施例中,将包含可疑人物的人脸的图像通过二值化、边缘提取及计算奇点等图像处理技术将可疑人物的面部特征,包括黑痣、胎记、鼻尖到双眼的距离、鼻梁的高度、长度及宽度、颧骨、双眼的距离、下颌轮廓、眉骨及其它具有明显识别作用的特征转化为特征点。
若由于图像模糊或面部偏转等原因造成所获取的特征点的数量小于设定的阈值,输出面部特征无效码。反之,若所获取的特征点的数量大于或等于设定的阈值,则输出面部特征码。其中,阈值为通过计算机多次训练得到的,在保证识别率情况下,计算量最小的特征点的数量。
手机中的面部特征数据库中存有大量目标人物的面部特征码,读取目标人物的面部特征码,并将输出的面部特征码与面部特征数据库中的目标人物的面部特征码进行一一匹配。匹配的结果可以反映为相似度,将[0%,100%]的相似度区间按相似度的大小划分为三个区间,匹配区间、待识别区间及不匹配区间。其中,匹配区间、待识别区间及不匹配区间可以设置在区间的边界连续或者不连续,如:将[0%,30%]划分为不匹配区间,[30%,70%]划分为待识别区间,[70%,100%]划分为匹配区间,或者将[0%,25%]划分为不匹配区间,[30%,70%]划分为待识别区间,[75%,100%]划分为匹配区间,具体的区间设置以实际要求为准。
若匹配得到的相似度属于匹配区间,则可疑人物为目标人物,处理器发出识别出目标人物的提示信息,并通过屏幕的文字、图案显示或声光报警提醒用户匹配成功;若匹配得到的相似度属于不匹配区间,处理器发出非目标人物的提示信息,并通过屏幕的文字、图案显示或声光报警提醒用户匹配成功,用户可以放弃对可疑人物的跟踪;若匹配得到的相似度属于待识别区间,表示不能确定可疑人物的身份,手机通过用户界面提示是否需要进一步采集可疑人物的图像。如果手机检测到用户发送了结束指令,或者因超时手机系统自动发出了结束指令,结束程序。否则,继续采集可疑人物的图像,直到手机发出识别出目标人物的提示信息或非目标人物的提示信息。
另外,手机获取的面部特征码可存储到面部特征数据库中,用于扩充面部特征数据库中的数据量,也可以通过基站与服务器连接,将获取的面部特征码通过无线网络传输到服务器中,更新服务器中的面部特征码数据。
另外,也可以在并将输出的面部特征码与面部特征数据库中的目标人物的面部特征码进行一一匹配前,向服务器发送请求更新数据库指令,并在接收到服务器发送的更新数据库指令后,接收服务器传输的面部特征码数据,并下载面部特征码数据以更新所述面部特征码数据库。
区别于现有技术的情况,当摄像头拍摄到的图像模糊不清或目标人物与摄像头偏转的角度超过限定的角度而造成匹配的结果位于待识别区间,即图像不能被识别但又不属于完全匹配或者完全不匹配的情况时,处理器会重新采集图像,直到图像能够准确被识别,提高识别率。并且,本发明结合人脸识别技术及移动终端技术,通过移动终端随时随地对目标人物进行采集图像,有效的扩大摄像头的视场。
参阅图3,本发明还提供了一种人脸识别装置,包括:
采集模块110:用于采集包含人脸在内的图像;
获取模块120:用于获取图像中人脸的面部特征;
判断模块130:用于在面部特征数据库中匹配获取的面部特征,判断匹配的结果属于匹配区间、不匹配区间及待识别区间的哪一个;
提示模块140:用于在判断属于匹配区间时,发出识别出目标人物的提示信息;在判断属于不匹配区间时,发出非目标人物的提示信息。
本发明的人脸识别装置可通过采集模块110采集一幅包含可疑人物的人脸在内的图像,并将该图像传输到获取模块120中提取可疑人物的面部特征,如黑痣、胎记、鼻尖到双眼的距离、鼻梁高度、长度、宽度、颧骨、双眼的距离、下颌轮廓、眉骨及其它具有明显识别作用的特征。
判断模块130从面部特征数据库中读取目标人物的面部特征,并将获取模块120输出的面部特征与面部特征数据库中的目标人物的面部特征进行一一匹配。匹配的结果可以反映为相似度,将[0%,100%]的相似度区间按相似度的大小划分为三个区间,匹配区间、待识别区间及不匹配区间。其中,匹配区间、待识别区间及不匹配区间可以设置在区间的边界连续或者不连续,如:将[0%,30%]划分为不匹配区间,[30%,70%]划分为待识别区间,[70%,100%]划分为匹配区间,或者将[0%,25%]划分为不匹配区间,[30%,70%]划分为待识别区间,[75%,100%]划分为匹配区间。具体的区间设置以实际要求为准。
若匹配得到的相似度属于匹配区间,则可疑人物为目标人物,提示模块140发出识别出目标人物的提示信息,并通过屏幕的文字、图案显示或声光报警提醒用户匹配成功;若匹配得到的相似度属于不匹配区间,提示模块140发出非目标人物的提示信息,并通过屏幕的文字、图案显示或声光报警提醒用户匹配成功,用户可以放弃对可疑人物的跟踪;若匹配得到的相似度属于待识别区间,表示不能确定可疑人物的身份,通过用户界面提示是否需要进一步采集可疑人物的图像。如果检测到用户发送了结束指令,或者因超时系统自动发出了结束指令,人脸识别装置停止工作。否则,采集模块110继续采集包含可疑人物的人脸的图像,直到提示模块140发出识别出目标人物的提示信息或非目标人物的提示信息。
区别于现有技术的情况,当摄像头拍摄到的图像模糊不清或目标人物与摄像头偏转的角度超过限定的角度而造成匹配的结果位于待识别区间,即图像不能被识别但又不属于完全匹配或者完全不匹配的情况时,处理器会重新采集图像,直到图像能够准确被识别,提高识别率。并且,本发明结合人脸识别技术及移动终端技术,通过移动终端随时随地对目标人物进行采集图像,有效的扩大摄像头的视场。
参阅图4,本发明还提供了一种能识别人脸的移动终端,下面依然以手机为例但不限于手机进行阐述,其中,手机的摄像头210的像素可以选择较高像素,以保证图像识别的要求:
发现可疑人物时,启动人脸识别程序,处理器230驱动LED灯变成红色,提示用户人脸识别程序处于运行中。此时,用户可通过手机的摄像头210采集一幅包含可疑人物的人脸的图像,并将该图像传输到识别模块220中。
识别模块220对图像采用直方图均衡化、图像平滑等方法进行处理得到清晰的图像。然后,对图像进行归一化处理,获得标准尺寸的图像。由于直方图均衡化、图像平滑及归一化均为现有技术,这里就不展开赘述。在得到清晰的图像后,识别模块220将包含可疑人物的人脸的图像通过二值化、边缘提取及计算奇点等图像处理技术将可疑人物的面部特征,如黑痣、胎记、鼻尖到双眼的距离、鼻梁高度、长度、宽度、颧骨、双眼的距离、下颌轮廓、眉骨及其它具有明显识别作用的特征,转化为特征点。若由于图像模糊或面部偏转等原因造成所获取的特征点的数量小于设定的阈值,识别模块220输出面部特征无效码。反之,若所获取的特征点的数量大于或等于设定的阈值,识别模块220输出面部特征码。其中,阈值为通过计算机多次训练得到的,在保证识别率情况下,计算量最小的特征点的数量。
处理器230接收到识别模块220输出的面部特征码后,从面部特征数据库中读取目标人物的面部特征码,并将识别模块220输出的面部特征与面部特征数据库中的目标人物的面部特征进行一一匹配。匹配的结果可以反映为相似度,将[0%,100%]的相似度区间按相似度的大小划分为三个区间,匹配区间、待识别区间及不匹配区间。其中,匹配区间、待识别区间及不匹配区间可以设置在区间的边界连续或者不连续,如:将[0%,30%]划分为不匹配区间,[30%,70%]划分为待识别区间,[70%,100%]划分为匹配区间,或者将[0%,25%]划分为不匹配区间,[30%,70%]划分为待识别区间,[75%,100%]划分为匹配区间。具体的区间设置以实际要求为准。
若匹配得到的相似度属于匹配区间,则可疑人物为目标人物,处理器230发出识别出目标人物的提示信息,并驱动LED灯变成绿色提醒用户匹配成功;若匹配得到的相似度属于不匹配区间,处理器230发出非目标人物的提示信息,并驱动LED灯变成绿色提醒用户匹配成功,用户可以放弃对可疑人物的跟踪;若匹配得到的相似度属于待识别区间,表示不能确定可疑人物的身份,处理器230驱动LED灯变成蓝色提醒用户需要进一步采集可疑人物的图像,手机通过用户界面提示是否需要进一步采集可疑人物的图像。如果手机检测到用户发送了结束指令,或者因超时手机系统自动发出了结束指令,结束程序,处理器230驱动LED灯熄灭。否则,继续采集可疑人物的图像,直到手机发出识别出目标人物的提示信息或非目标人物的提示信息。
另外,手机获取的面部特征码可存储到本机的面部特征数据库中,用于扩充面部特征数据库中的数据量,也可以通过射频模块240与服务器连接,将获取的面部特征码通过无线网络传输到服务器中,更新服务器中的面部特征码数据。
另外,也可以在并将输出的面部特征码与面部特征数据库中的目标人物的面部特征码进行一一匹配前,通过射频模块240向服务器发送请求更新数据库指令,并在接收到服务器发送的更新数据库指令后,接收服务器传输的面部特征码数据,并下载面部特征码数据以更新面部特征码数据库。射频模块240可以是GSM、GPRS、TDS-CDMA、CDMA2000、WCDMA中的一种或多种。
区别于现有技术的情况,当摄像头210拍摄到的图像模糊不清或目标人物与摄像头210偏转的角度超过限定的角度而造成匹配的结果位于待识别区间,即图像不能被识别时,处理器230会重新采集图像,直到图像能够准确被识别。并且,本发明结合人脸识别技术及移动终端技术,通过移动终端随时随地对目标人物进行采集图像,有效的扩大摄像头的视场。
以上所述仅为本发明的实施例,并非因此限制本发明的专利范围,凡是利用本发明说明书及附图内容所作的等效结构或等效流程变换,或直接或间接运用在其他相关的技术领域,均同理包括在本发明的专利保护范围内。

Claims (12)

  1. 一种能识别人脸的移动终端,其中,所述移动终端包括:
    摄像头,用于采集包含人脸在内的图像;
    识别模块,用于获取图像中人脸的面部特征点,若特征点的数量小于设定的阈值,则输出面部特征无效码;若特征点的数量大于或等于设定的阈值,则输出面部特征码;
    处理器,用于在面部特征数据库中匹配输出的面部特征码,判断所述输出的面部特征码与面部特征码数据库中的面部特征码的相似度属于匹配区间、不匹配区间及待识别区间的哪一个;若判断属于匹配区间,所述处理器发出识别出目标人物的提示信息,若判断属于不匹配区间,所述处理器发出非目标人物的提示信息,若判断属于待识别区间,所述处理器控制所述摄像头重新采集包含人脸在内的图像,直到所述处理器检测到结束指令或发出识别出目标人物的提示信息或非目标人物的提示信息,以及在启动人脸识别时,驱动LED灯变成红色;在匹配结果属于匹配区间或不匹配区间时,所述驱动LED灯变成绿色;在匹配结果属于待识别区间时,所述驱动LED灯变成蓝色;在检测到结束命令时,所述驱动LED灯熄灭。
  2. 根据权利要求1所述的移动终端,其中,所述移动终端还包括:
    射频模块,用于向服务器发送请求更新数据库指令及在接收到服务器发送的更新数据库指令时,接收服务器传输的数据。
  3. 根据权利要求2所述的移动终端,其中,所述射频模块还用于向所述服务器传输面部特征码数据以更新所述服务器中的面部特征码数据。
  4. 一种人脸识别方法,其中,包括:
    采集包含人脸在内的图像;
    获取所述图像中人脸的面部特征;
    在面部特征数据库中匹配所述获取的面部特征,判断匹配的结果属于匹配区间、不匹配区间及待识别区间的哪一个;
    若判断属于匹配区间,发出识别出目标人物的提示信息;若判断属于不匹配区间,发出非目标人物的提示信息;若判断属于待识别区间,返回执行采集包含人脸在内的图像的步骤,直到检测到结束指令或发出识别出目标人物的提示信息或非目标人物的提示信息。
  5. 根据权利要求4所述的人脸识别方法,其中,所述获取图像中人脸的面部特征的步骤包括:获取图像中人脸的面部特征点;
    在所述面部特征数据库中匹配获取的面部特征的步骤之后,包括:若特征点的数量小于设定的阈值,则输出面部特征无效码;若特征点的数量大于或等于设定的阈值,则输出面部特征码;
    所述在面部特征数据库中匹配获取的面部特征的步骤包括:在面部特征数据库中匹配输出的面部特征码;
    所述判断匹配的结果为属于匹配区间、不匹配区间及待识别区间的哪一个的步骤包括:判断所述输出的面部特征码与面部特征码数据库中的面部特征码的相似度属于匹配区间、不匹配区间及待识别区间的哪一个。
  6. 根据权利要求5所述的人脸识别方法,其中,所述在面部特征数据库中匹配输出的面部特征码步骤之前还包括如下步骤:
    向服务器发送请求更新数据库指令;
    若接收到服务器发送的更新数据库指令,接收服务器传输的面部特征码数据,并下载所述面部特征码数据以更新所述面部特征码数据库。
  7. 根据权利要求5所述的人脸识别方法,其中,所述在特征点的数量大于或等于设定的阈值时,输出面部特征码步骤之后还包括如下步骤:
    向所述服务器传输面部特征码数据以更新所述服务器中的面部特征码数据。
  8. 一种能识别人脸的移动终端,其中,所述移动终端包括:
    摄像头,用于采集包含人脸在内的图像;
    识别模块,用于获取所述图像中人脸的面部特征;
    处理器,用于在面部特征数据库中匹配所述获取的面部特征,判断匹配的结果属于匹配区间、不匹配区间及待识别区间的哪一个;
    若判断属于匹配区间,所述处理器发出识别出目标人物的提示信息,若判断属于不匹配区间,所述处理器发出非目标人物的提示信息,若判断属于待识别区间,所述处理器控制所述摄像头重新采集包含人脸在内的图像,直到所述处理器检测到结束指令或发出识别出目标人物的提示信息或非目标人物的提示信息。
  9. 根据权利要求8所述的移动终端,其中,
    所述识别模块还进一步用于获取图像中人脸的面部特征点,若特征点的数量小于设定的阈值,则输出面部特征无效码;若特征点的数量大于或等于设定的阈值,则输出面部特征码;
    所述处理器还进一步用于在面部特征数据库中匹配输出的面部特征码,判断所述输出的面部特征码与面部特征码数据库中的面部特征码的相似度属于匹配区间、不匹配区间及待识别区间的哪一个。
  10. 根据权利要求9所述的移动终端,其中,所述移动终端还包括:
    射频模块,用于向服务器发送请求更新数据库指令及在接收到服务器发送的更新数据库指令时,接收服务器传输的数据。
  11. 根据权利要求10所述的移动终端,其中,所述射频模块还用于向所述服务器传输面部特征码数据以更新所述服务器中的面部特征码数据。
  12. 根据权利要求8所述的移动终端,其中,
    所述处理器还用于在启动人脸识别时,驱动LED灯变成红色;在匹配结果属于匹配区间或不匹配区间时,所述驱动LED灯变成绿色;在匹配结果属于待识别区间时,所述驱动LED灯变成蓝色;在检测到结束命令时,所述驱动LED灯熄灭。
PCT/CN2012/081281 2011-12-07 2012-09-12 人脸识别方法及能识别人脸的移动终端 WO2013082957A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201110404997.3A CN102542299B (zh) 2011-12-07 2011-12-07 人脸识别方法、装置及能识别人脸的移动终端
CN201110404997.3 2011-12-07

Publications (1)

Publication Number Publication Date
WO2013082957A1 true WO2013082957A1 (zh) 2013-06-13

Family

ID=46349149

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2012/081281 WO2013082957A1 (zh) 2011-12-07 2012-09-12 人脸识别方法及能识别人脸的移动终端

Country Status (2)

Country Link
CN (1) CN102542299B (zh)
WO (1) WO2013082957A1 (zh)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109117821A (zh) * 2018-08-31 2019-01-01 广东九联科技股份有限公司 一种执法记录仪的人脸识别系统及方法
CN109658085A (zh) * 2018-09-10 2019-04-19 杭州乐优客科技有限公司 一种刷脸识别会员的收银系统
CN110688901A (zh) * 2019-08-26 2020-01-14 苏宁云计算有限公司 一种人脸识别方法及装置
CN110782241A (zh) * 2019-10-31 2020-02-11 郑州天迈科技股份有限公司 一种混合人脸识别公交支付系统及方法
CN111444899A (zh) * 2020-05-14 2020-07-24 聚好看科技股份有限公司 远程考试控制方法、服务器及终端
CN111797746A (zh) * 2020-06-28 2020-10-20 北京小米松果电子有限公司 人脸识别方法、装置及计算机可读存储介质
CN112180621A (zh) * 2019-07-04 2021-01-05 奇酷互联网络科技(深圳)有限公司 调节眼镜度数的方法、终端和存储介质
CN113003339A (zh) * 2021-02-22 2021-06-22 上海三菱电梯有限公司 电梯识别方法、识别系统与电梯
CN113111780A (zh) * 2021-04-13 2021-07-13 谢爱菊 一种基于区块链的区域报警监管系统及方法
CN113473075A (zh) * 2020-07-14 2021-10-01 青岛海信电子产业控股股份有限公司 一种视频监控数据隐私保护方法及其装置
CN113920752A (zh) * 2021-09-28 2022-01-11 云南思立耐特信息系统工程有限公司 一种智能交通基础设施监控管理系统

Families Citing this family (34)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102542299B (zh) * 2011-12-07 2015-03-25 惠州Tcl移动通信有限公司 人脸识别方法、装置及能识别人脸的移动终端
CN103268551A (zh) * 2013-05-29 2013-08-28 苏州福丰科技有限公司 基于人脸识别的银行卡支付系统及其方法
CN104680053B (zh) * 2013-12-03 2018-05-11 湖北海洋文化传播有限公司 对当前认证终端持有人身份认证的方法及装置
CN104780308A (zh) * 2014-01-09 2015-07-15 联想(北京)有限公司 一种信息处理方法及电子设备
CN104541230B (zh) * 2014-11-15 2017-07-04 深圳市三木智能技术有限公司 视频验证防止手机误操作的方法及装置
CN105447474B (zh) * 2015-12-15 2019-08-06 3M创新有限公司 人像识别方法和人像识别装置
CN108304827A (zh) * 2016-04-01 2018-07-20 张海东 自动化面部识别系统
CN105868613A (zh) * 2016-06-08 2016-08-17 广东欧珀移动通信有限公司 一种生物特征识别方法、装置和移动终端
CN106295504A (zh) * 2016-07-26 2017-01-04 车广为 人脸识别基础上的增强显示方法
CN106547831A (zh) * 2016-10-10 2017-03-29 天脉聚源(北京)科技有限公司 提高图像信息的处理效率的方法及装置
CN108573268A (zh) * 2017-03-10 2018-09-25 北京旷视科技有限公司 图像识别方法和装置、图像处理方法和装置及存储介质
CN107578006B (zh) * 2017-08-31 2020-06-23 维沃移动通信有限公司 一种照片处理方法及移动终端
CN107679559B (zh) * 2017-09-15 2020-01-10 Oppo广东移动通信有限公司 图像处理方法、装置、计算机可读存储介质和移动终端
CN107808081B (zh) * 2017-10-12 2020-01-14 Oppo广东移动通信有限公司 一种提醒方法及相关设备
CN107832692A (zh) * 2017-10-31 2018-03-23 浪潮金融信息技术有限公司 一种人脸图像处理方法和装置
CN109308754A (zh) * 2017-11-16 2019-02-05 中国石油化工股份有限公司 基于人脸识别的外来作业人员身份识别装置
CN108305362A (zh) * 2018-01-23 2018-07-20 吴君国 安防监控设备集中控制系统
CN110609562B (zh) * 2018-06-15 2021-07-16 华为技术有限公司 一种图像信息采集方法和装置
CN108924426B (zh) * 2018-08-01 2020-01-10 Oppo广东移动通信有限公司 图像处理方法和装置、电子设备、计算机可读存储介质
CN109120846B (zh) * 2018-08-01 2020-01-10 Oppo广东移动通信有限公司 图像处理方法和装置、电子设备、计算机可读存储介质
WO2020024575A1 (zh) 2018-08-01 2020-02-06 Oppo广东移动通信有限公司 图像处理方法和装置、电子设备、计算机可读存储介质
CN109327678A (zh) * 2018-09-10 2019-02-12 国网山东省电力公司淄博供电公司 一种基于人脸识别摄像头的机房内作业区域管控方法
CN109447006A (zh) * 2018-11-01 2019-03-08 北京旷视科技有限公司 图像处理方法、装置、设备及存储介质
CN109858464B (zh) * 2019-02-26 2021-03-23 北京旷视科技有限公司 底库数据处理方法、人脸识别方法、装置和电子设备
CN109951637B (zh) * 2019-03-19 2020-09-11 河北川谷信息技术有限公司 基于大数据的安防监控探头分析处理方法
CN109856800B (zh) * 2019-04-09 2021-09-03 中科海微(北京)科技有限公司 分体式ar眼镜的显示控制方法、装置及分体式ar眼镜
CN110298667A (zh) * 2019-06-25 2019-10-01 北京百度网讯科技有限公司 基于人脸识别的支付处理方法、装置及存储介质
CN113095110B (zh) * 2019-12-23 2024-03-08 浙江宇视科技有限公司 人脸数据动态入库的方法、装置、介质及电子设备
CN111241930A (zh) * 2019-12-30 2020-06-05 航天信息股份有限公司 一种用于人脸识别的方法及系统
CN111144332A (zh) * 2019-12-30 2020-05-12 深圳云天励飞技术有限公司 一种图片聚档方法、装置和电子设备
CN113139407A (zh) * 2020-01-20 2021-07-20 深圳市道控技术有限公司 一种人脸面部识别控制方法、装置及存储介质
CN114945102A (zh) * 2020-07-14 2022-08-26 海信视像科技股份有限公司 显示设备及人物识别展示的方法
CN114312666A (zh) * 2021-11-22 2022-04-12 江铃汽车股份有限公司 基于人脸识别的车辆控制方法、装置、存储介质及设备
CN113901423B (zh) * 2021-12-08 2022-09-13 云丁网络技术(北京)有限公司 一种基于人脸识别的智能安防设备控制方法和系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101216881A (zh) * 2007-12-28 2008-07-09 北京中星微电子有限公司 一种图像自动获取方法和装置
JP4155875B2 (ja) * 2003-06-16 2008-09-24 シャープ株式会社 画像撮像装置
CN101440676A (zh) * 2008-12-22 2009-05-27 北京中星微电子有限公司 一种基于摄像头的智能防盗门锁及其报警处理方法
CN102542299A (zh) * 2011-12-07 2012-07-04 惠州Tcl移动通信有限公司 人脸识别方法、装置及能识别人脸的移动终端

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100561502C (zh) * 2007-12-18 2009-11-18 北京中星微电子有限公司 一种人脸认证的方法和装置
CN102073842A (zh) * 2009-11-25 2011-05-25 天津三星光电子有限公司 一种通过分类匹配的人脸识别方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4155875B2 (ja) * 2003-06-16 2008-09-24 シャープ株式会社 画像撮像装置
CN101216881A (zh) * 2007-12-28 2008-07-09 北京中星微电子有限公司 一种图像自动获取方法和装置
CN101440676A (zh) * 2008-12-22 2009-05-27 北京中星微电子有限公司 一种基于摄像头的智能防盗门锁及其报警处理方法
CN102542299A (zh) * 2011-12-07 2012-07-04 惠州Tcl移动通信有限公司 人脸识别方法、装置及能识别人脸的移动终端

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109117821A (zh) * 2018-08-31 2019-01-01 广东九联科技股份有限公司 一种执法记录仪的人脸识别系统及方法
CN109658085A (zh) * 2018-09-10 2019-04-19 杭州乐优客科技有限公司 一种刷脸识别会员的收银系统
CN109658085B (zh) * 2018-09-10 2023-12-05 全通金信控股(广东)有限公司 一种刷脸识别会员的收银系统
CN112180621A (zh) * 2019-07-04 2021-01-05 奇酷互联网络科技(深圳)有限公司 调节眼镜度数的方法、终端和存储介质
CN110688901A (zh) * 2019-08-26 2020-01-14 苏宁云计算有限公司 一种人脸识别方法及装置
CN110782241A (zh) * 2019-10-31 2020-02-11 郑州天迈科技股份有限公司 一种混合人脸识别公交支付系统及方法
CN110782241B (zh) * 2019-10-31 2024-03-29 郑州天迈科技股份有限公司 一种混合人脸识别公交支付系统及方法
CN111444899A (zh) * 2020-05-14 2020-07-24 聚好看科技股份有限公司 远程考试控制方法、服务器及终端
CN111444899B (zh) * 2020-05-14 2023-10-31 聚好看科技股份有限公司 远程考试控制方法、服务器及终端
CN111797746A (zh) * 2020-06-28 2020-10-20 北京小米松果电子有限公司 人脸识别方法、装置及计算机可读存储介质
CN113473075A (zh) * 2020-07-14 2021-10-01 青岛海信电子产业控股股份有限公司 一种视频监控数据隐私保护方法及其装置
CN113003339B (zh) * 2021-02-22 2022-12-20 上海三菱电梯有限公司 电梯识别方法、识别系统与电梯
CN115490104A (zh) * 2021-02-22 2022-12-20 上海三菱电梯有限公司 电梯识别方法、识别系统与电梯
CN115490104B (zh) * 2021-02-22 2023-10-17 上海三菱电梯有限公司 电梯识别方法、识别系统与电梯
CN113003339A (zh) * 2021-02-22 2021-06-22 上海三菱电梯有限公司 电梯识别方法、识别系统与电梯
CN113111780A (zh) * 2021-04-13 2021-07-13 谢爱菊 一种基于区块链的区域报警监管系统及方法
CN113920752A (zh) * 2021-09-28 2022-01-11 云南思立耐特信息系统工程有限公司 一种智能交通基础设施监控管理系统

Also Published As

Publication number Publication date
CN102542299B (zh) 2015-03-25
CN102542299A (zh) 2012-07-04

Similar Documents

Publication Publication Date Title
WO2013082957A1 (zh) 人脸识别方法及能识别人脸的移动终端
WO2017156964A1 (zh) 一种终端解锁的方法及终端
WO2010008143A2 (ko) 생체지문 판단장치 및 그 판단방법
CN105224906B (zh) 掌静脉识别智能系统
CN105608649A (zh) 一种基于人脸识别的酒店入住管理系统
CN105975938A (zh) 一种智能动态人脸识别社区管家服务系统
CN204791017U (zh) 基于多生物特征模态的移动终端用户身份认证装置
WO2018072099A1 (zh) 一种手掌检测方法和设备
CN104966061A (zh) 一种掌纹识别方法及装置、智能安全门
CN110097006A (zh) 一种基于手部掌纹掌脉特征融合的非接触式生物识别方法
CN205540806U (zh) 一种三维人脸图像识别装置
WO2019104551A1 (zh) 指纹识别方法及终端设备
CN108985211B (zh) 一种刑事侦查用指纹采集系统
WO2020073169A1 (zh) 生物特征识别的方法、装置和电子设备
CN107665328A (zh) 人脸面部识别方法
CN103825921A (zh) 基于网络传输的生物信息交互系统及其实现方法
CN208335306U (zh) 识别终端和管理系统
CN213659511U (zh) 基于人脸与指静脉融合识别的考勤设备及考勤系统
CN206672129U (zh) 门禁系统的指纹识别器
CN207216727U (zh) 一种访客识别管理系统
CN202058170U (zh) 一种结合摄像头和指纹识别器的活体指纹识别装置
CN112465999A (zh) 基于人脸与指静脉融合识别的考勤设备及考勤系统
WO2021184690A1 (zh) 基于视觉识别的传染疾病检测机器人及其检测方法
CN210052203U (zh) 一种基于多种识别方法的考勤签到系统
CN202771442U (zh) 便携式人脸识别设备

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: 12855415

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: 12855415

Country of ref document: EP

Kind code of ref document: A1