CN110874588A - Method and device for dynamically optimizing light influence in face recognition - Google Patents
Method and device for dynamically optimizing light influence in face recognition Download PDFInfo
- Publication number
- CN110874588A CN110874588A CN202010051299.9A CN202010051299A CN110874588A CN 110874588 A CN110874588 A CN 110874588A CN 202010051299 A CN202010051299 A CN 202010051299A CN 110874588 A CN110874588 A CN 110874588A
- Authority
- CN
- China
- Prior art keywords
- face
- infrared
- registry
- photo
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/50—Maintenance of biometric data or enrolment thereof
Abstract
The invention relates to the technical field of image processing, and provides a method and a system for dynamically optimizing light influence in face recognition, which comprises the following steps: simultaneously acquiring data of the two cameras through the binocular cameras, and extracting face characteristic values; and preferentially using the infrared camera data to perform face detection processing, retrieving the infrared face registry for identification, entering the bottom library photo optimization of the infrared face registry if successful matching is performed, and entering the face detection processing of the RGB camera data if unsuccessful matching is performed. The method and the system for dynamically optimizing the light influence in the face recognition use the infrared camera to perform the whole-process face recognition processing, so that the face recognition processing is basically not influenced by external light conditions any more, and the face recognition technology and the environmental robustness of the application in various fields can be greatly improved.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a method for dynamically optimizing light influence in face recognition.
Background
The face recognition comprises face detection, living body detection, feature extraction and face comparison, and finally the whole process of the matching result and corresponding face recognition application are displayed through threshold judgment. Before the process is used, a face library needs to be registered in advance for retrieval and identification in the face library when the face is identified. Because the imaging of the RGB camera is greatly influenced by the light environment, the human face detection, the living body detection and the human face comparison result become very poor under the condition of poor light conditions.
In order to solve the problems, the prior art is improved to some extent from two aspects, one is to upgrade an RGB camera to a camera supporting wide dynamics, so that the imaging effect adjusting capability in a poor light scene can be enhanced to some extent, and the imaging effect is improved; and secondly, a binocular camera is used, and an infrared imaging camera is additionally arranged on the basis of the RGB camera and used for enhancing the usability of the living body detection under the condition of poor light. The infrared lens is used for imaging, and the imaging is performed by using the irradiation of an infrared light source which actively emits 850/940nm (different infrared wave bands are selected according to different irradiation requirements in daily use), so that the infrared lens is hardly influenced by external light, and very stable imaging quality can be obtained.
However, the above improvements, while better solving the usability problem of live body detection under poor lighting conditions, do not solve the adverse effects of poor lighting conditions on face detection, feature extraction and face comparison based on RGB camera data. In terms of the overall face recognition effect, adverse influence factors of light conditions in face recognition are not eliminated to the maximum extent.
In the existing face recognition application, the registration of the infrared photos can only be directly registered at the device end provided with the infrared camera, the registration efficiency is low, and the actual application of the large-scale infrared face library hardly has practical feasibility. This is one of the reasons why data generated by the infrared camera cannot be generally applied to face recognition processing other than live body detection at present.
Disclosure of Invention
The invention aims to provide a method for dynamically optimizing light influence in face recognition, which uses an infrared camera to perform whole-process face recognition processing, can ensure that the face recognition processing is basically not influenced by external light conditions any more, and can greatly improve the face recognition technology and the environmental robustness applied in various fields.
To achieve the above object, the present invention provides a method for dynamically optimizing the influence of light in face recognition, comprising:
step 1, simultaneously acquiring data of two cameras through a binocular camera with an RGB camera and an infrared camera, and extracting a face characteristic value after a face is detected;
step 2, preferentially using the infrared camera data to perform face detection processing, retrieving an infrared face registry to perform face recognition, entering the bottom library photo optimization of the infrared face registry if the face features in the infrared face registry are successfully matched, and performing face data re-acquisition or entering the face detection processing of the data of the RGB camera according to whether the data of the RGB camera detects the face if the matching is unsuccessful;
the method comprises the steps that a base photo of an infrared face registry is optimized, wherein the base photo optimization of the infrared face registry comprises the step of dynamically replacing an infrared face photo which is registered in the infrared face registry by using a currently obtained infrared face photo;
the face detection processing of the data of the RGB camera comprises the steps of responding to the face characteristics successfully matched into the RGB face registry, performing infrared photo quality detection, and adding and registering infrared face data corresponding to the person to the infrared face registry when the quality of the infrared face photos currently and simultaneously acquired by the RGB camera is higher than that of the person in the infrared face registry.
Preferably, in the step 2, the processing of optimizing the base library photos of the infrared face registry comprises:
calling an availability statistical module of the infrared face registry, detecting the registered infrared face photos, and returning a result of whether the photo optimization of the infrared face registry is needed:
if the registered infrared face photos in the infrared face registry do not need to be updated, the display module directly displays the successful recognition;
if the photo of the infrared face registry is required to be optimized, an infrared face photo quality evaluation module is called to detect, when the quality of the currently acquired infrared face photo is higher than that of the infrared face photo of the person in the infrared face registry, dynamic replacement is carried out, and the currently acquired infrared face photo is used for dynamically replacing the infrared face photo registered in the infrared face registry; and after the infrared registration photo is optimized, the display module displays that the identification is successful.
Preferably, in the step 2, it is determined whether the data of the RGB camera detects a face, if not, the step 1 is returned to perform the re-acquisition of the face data, otherwise, the face detection processing of the data of the RGB camera is performed.
Preferably, the face detection processing of the data of the RGB camera includes: and if the face features in the RGB face registry are not successfully matched, returning to the step 1 to carry out the re-acquisition of the face data.
Preferably, in the step 2, the infrared picture quality detection includes:
and (3) calling an infrared human face photo quality evaluation module for detection, and returning to the step (1) for re-acquisition of human face data if the quality of the infrared human face photos currently acquired simultaneously with the RGB camera is lower than that of the human face photos of the person in the infrared human face registry.
Preferably, after the infrared face data corresponding to the registered person is added to the infrared face registry, the first automatic registration processing of the infrared registered photo of the person is completed, and the display module displays that the recognition is successful.
According to the present invention, there is also provided a system for dynamically optimizing the influence of light in face recognition, comprising:
the binocular camera arranged on the face recognition equipment comprises an RGB (red, green and blue) camera and an infrared camera, when a person passes through the binocular camera, data of the two cameras are simultaneously acquired through the RGB camera and the infrared camera, and after a face is detected, a face characteristic value is extracted;
at least one computing device, configured to preferentially use the infrared camera data to perform face detection processing, retrieve the infrared face registry to perform face recognition, enter the base photo optimization of the infrared face registry if the face features successfully matched in the infrared face registry, and perform face data re-acquisition or enter the face detection processing of the RGB camera data according to whether the data of the RGB camera detects a face if the matching is unsuccessful;
the method comprises the steps that a base photo of an infrared face registry is optimized, wherein the base photo optimization of the infrared face registry comprises the step of dynamically replacing an infrared face photo which is registered in the infrared face registry by using a currently obtained infrared face photo;
the face detection processing of the data of the RGB camera comprises the steps of responding to the face characteristics successfully matched into the RGB face registry, performing infrared photo quality detection, and adding and registering infrared face data corresponding to the person to the infrared face registry when the quality of the infrared face photos currently and simultaneously acquired by the RGB camera is higher than that of the person in the infrared face registry.
Preferably, the computing device is further arranged to:
calling an availability statistical module of the infrared face registry, detecting the registered infrared face photos, and returning a result of whether the photo optimization of the infrared face registry is needed:
if the registered infrared face photos in the infrared face registry do not need to be updated, the display module directly displays the successful recognition;
if the photo of the infrared face registry is required to be optimized, an infrared face photo quality evaluation module is called to detect, when the quality of the currently acquired infrared face photo is higher than that of the infrared face photo of the person in the infrared face registry, dynamic replacement is carried out, and the currently acquired infrared face photo is used for dynamically replacing the infrared face photo registered in the infrared face registry; and after the infrared registration photo is optimized, the display module displays that the identification is successful.
Preferably, the at least one computing device is further configured to:
and after the infrared face data corresponding to the registered person is added to the infrared face registry, the first automatic registration processing of the infrared registered photo of the person is completed, and the successful recognition is displayed.
Preferably, the binocular camera is provided on the at least one computing device.
Thus, in a specific operation, the method for dynamically optimizing the light influence in the face recognition of the present invention operates as follows:
step 1, setting binocular cameras, namely an RGB (red, green and blue) camera and an infrared camera on face recognition equipment, simultaneously acquiring data of the two cameras, and preferentially performing face detection processing on the infrared camera data; after a face is detected, extracting a face characteristic value;
step 2, using the face characteristic value corresponding to the acquired infrared camera data, preferentially searching an infrared face registry, performing face recognition, if the face characteristic in the infrared face registry is successfully matched, calling an availability statistical module of the infrared face registry, detecting a registered infrared face photo, returning whether the photo optimization of the infrared face registry is needed, and if the registered infrared face photo in the infrared face registry is not needed to be updated, directly displaying the successful recognition by a display module; if the photo optimization of the infrared face registry is needed, entering the step 3;
step 3, calculating by using an infrared human face photo quality evaluation module, dynamically replacing the currently acquired infrared human face photo when the quality of the infrared human face photo is higher than that of the person in the infrared human face registration library, dynamically replacing the currently acquired infrared human face photo with higher-quality infrared human face registration data, and displaying and identifying successfully by using a display module after the optimization processing of the infrared human face photo is completed;
step 4, if no person can be identified and the RGB camera cannot detect the face in step 1 when the face characteristic value corresponding to the infrared camera data is used for retrieving the infrared face registry in step 2, returning to step 1 to restart a new face detection attempt, and if the RGB camera successfully detects the face in step 1, entering step 2 to further use the face characteristic value corresponding to the RGB camera data to retrieve the RGB face registry;
and 5, if the RGB face library retrieval processing in the step 4 fails to identify any person, the person cannot be effectively identified by the identification, and the step 1 is returned to restart a new face detection attempt. If the RGB face library retrieval processing in step 4 successfully matches the face features (higher than the preset hit threshold), an infrared face picture quality evaluation module is invoked to perform calculation, and when the quality of the current infrared face picture acquired simultaneously with the RGB camera is higher than the picture quality of the person in the infrared face registry, infrared face data corresponding to the person is added to the infrared face registry (this key processing completes the initial automatic registration of the infrared face registration data of the person originally having only RGB face registration data). And after the automatic registration processing of the infrared registration photo of the person is completed, the display module displays that the identification is successful.
Therefore, the invention can automatically check whether the corresponding personnel has infrared face registration data after successfully identifying the personnel in the RGB face registry. If no infrared face registration data exists, automatically registering the infrared face data which currently meets the registration quality requirement after the infrared face imaging quality is automatically evaluated; and after the corresponding personnel of the infrared face registry is successfully identified, the availability evaluation of the infrared face registry photos of the personnel is automatically carried out. If the infrared face registration photo of the person needs to be optimized in evaluation, after the infrared face imaging quality is automatically evaluated, the infrared face registration data of the person is automatically updated by using the infrared face data which currently meets the registration quality requirement and has quality superior to that of the registered infrared face data of the person.
Drawings
Fig. 1 is a schematic diagram of a method for dynamically optimizing the influence of light in face recognition according to the present invention.
Detailed Description
In order to better understand the technical content of the present invention, specific embodiments are described below with reference to the accompanying drawings.
With reference to the exemplary flow shown in fig. 1, the method for dynamically optimizing the light influence in face recognition according to the present invention includes the following steps:
step 1, simultaneously acquiring data of two cameras through a binocular camera with an RGB camera and an infrared camera, and extracting a face characteristic value after a face is detected;
step 2, preferentially using the infrared camera data to perform face detection processing, retrieving an infrared face registry to perform face recognition, entering the bottom library photo optimization of the infrared face registry if the face features in the infrared face registry are successfully matched, and performing face data re-acquisition or entering the face detection processing of the data of the RGB camera according to whether the data of the RGB camera detects the face if the matching is unsuccessful;
the method comprises the steps that a base photo of an infrared face registry is optimized, wherein the base photo optimization of the infrared face registry comprises the step of dynamically replacing an infrared face photo which is registered in the infrared face registry by using a currently obtained infrared face photo;
the face detection processing of the data of the RGB camera comprises the steps of responding to the face characteristics successfully matched into the RGB face registry, performing infrared photo quality detection, and adding and registering infrared face data corresponding to the person to the infrared face registry when the quality of the infrared face photos currently and simultaneously acquired by the RGB camera is higher than that of the person in the infrared face registry.
Preferably, in the step 2, the processing of optimizing the base library photos of the infrared face registry comprises:
calling an availability statistical module of the infrared face registry, detecting the registered infrared face photos, and returning a result of whether the photo optimization of the infrared face registry is needed:
if the registered infrared face photos in the infrared face registry do not need to be updated, the display module directly displays the successful recognition;
if the photo of the infrared face registry is required to be optimized, an infrared face photo quality evaluation module is called to detect, when the quality of the currently acquired infrared face photo is higher than that of the infrared face photo of the person in the infrared face registry, dynamic replacement is carried out, and the currently acquired infrared face photo is used for dynamically replacing the infrared face photo registered in the infrared face registry; and after the infrared registration photo is optimized, the display module displays that the identification is successful.
Preferably, in the step 2, it is determined whether the data of the RGB camera detects a face, if not, the step 1 is returned to perform the re-acquisition of the face data, otherwise, the face detection processing of the data of the RGB camera is performed.
Preferably, the face detection processing of the data of the RGB camera includes: and if the face features in the RGB face registry are not successfully matched, returning to the step 1 to carry out the re-acquisition of the face data.
Preferably, in the step 2, the infrared picture quality detection includes:
and (3) calling an infrared human face photo quality evaluation module for detection, and returning to the step (1) for re-acquisition of human face data if the quality of the infrared human face photos currently acquired simultaneously with the RGB camera is lower than that of the human face photos of the person in the infrared human face registry.
Preferably, after the infrared face data corresponding to the registered person is added to the infrared face registry, the first automatic registration processing of the infrared registered photo of the person is completed, and the display module displays that the recognition is successful.
According to the present invention, there is also provided a system for dynamically optimizing the influence of light in face recognition, comprising:
the binocular camera arranged on the face recognition equipment comprises an RGB (red, green and blue) camera and an infrared camera, when a person passes through the binocular camera, data of the two cameras are simultaneously acquired through the RGB camera and the infrared camera, and after a face is detected, a face characteristic value is extracted;
at least one computing device, configured to preferentially use the infrared camera data to perform face detection processing, retrieve the infrared face registry to perform face recognition, enter the base photo optimization of the infrared face registry if the face features successfully matched in the infrared face registry, and perform face data re-acquisition or enter the face detection processing of the RGB camera data according to whether the data of the RGB camera detects a face if the matching is unsuccessful;
the method comprises the steps that a base photo of an infrared face registry is optimized, wherein the base photo optimization of the infrared face registry comprises the step of dynamically replacing an infrared face photo which is registered in the infrared face registry by using a currently obtained infrared face photo;
the face detection processing of the data of the RGB camera comprises the steps of responding to the face characteristics successfully matched into the RGB face registry, performing infrared photo quality detection, and adding and registering infrared face data corresponding to the person to the infrared face registry when the quality of the infrared face photos currently and simultaneously acquired by the RGB camera is higher than that of the person in the infrared face registry.
Preferably, the computing device is further arranged to:
calling an availability statistical module of the infrared face registry, detecting the registered infrared face photos, and returning a result of whether the photo optimization of the infrared face registry is needed:
if the registered infrared face photos in the infrared face registry do not need to be updated, the display module directly displays the successful recognition;
if the photo of the infrared face registry is required to be optimized, an infrared face photo quality evaluation module is called to detect, when the quality of the currently acquired infrared face photo is higher than that of the infrared face photo of the person in the infrared face registry, dynamic replacement is carried out, and the currently acquired infrared face photo is used for dynamically replacing the infrared face photo registered in the infrared face registry; and after the infrared registration photo is optimized, the display module displays that the identification is successful.
Preferably, the at least one computing device is further configured to:
and after the infrared face data corresponding to the registered person is added to the infrared face registry, the first automatic registration processing of the infrared registered photo of the person is completed, and the successful recognition is displayed.
Preferably, the binocular camera is provided on the at least one computing device.
Thus, in a specific operation, the method for dynamically optimizing the light influence in the face recognition of the present invention operates as follows:
step 1, setting binocular cameras, namely an RGB (red, green and blue) camera and an infrared camera on face recognition equipment, simultaneously acquiring data of the two cameras, and preferentially performing face detection processing on the infrared camera data; after a face is detected, extracting a face characteristic value; the RGB camera may not detect the human face due to poor light conditions, but does not influence the human face recognition process of the method; (if the subsequent processing interruption can directly return to the step, and the current latest imaging data of the two cameras are obtained again for processing)
Step 2, using the face characteristic value corresponding to the acquired infrared camera data, preferentially searching an infrared face registry, and performing face recognition, if the face characteristic in the infrared face registry is successfully matched (namely higher than a preset hit threshold), calling an availability statistical module of the infrared face registry, detecting registered infrared face photos, returning whether the photo optimization of the infrared face registry is needed, and if the registered infrared face photos in the infrared face registry are not needed to be updated, directly displaying the successful recognition (the person is successfully recognized through the face data acquired by the infrared camera) by a display module; if the photo optimization of the infrared face registry is needed, entering the step 3;
step 3, calculating by using an infrared human face photo quality evaluation module, dynamically replacing (completing automatic optimization of registered infrared human face registration data of the person, dynamically replacing the registered infrared human face registration data with higher quality to further realize better infrared recognition passing effect) when the quality of the currently acquired infrared human face photo is higher than that of the person in an infrared human face registration library, dynamically replacing the registered infrared human face photo with higher quality infrared human face registration data, and displaying the registered infrared human face photo with successful recognition by a display module after the optimization processing of the infrared human face photo is completed;
step 4, if no person can be identified when the face characteristic value corresponding to the infrared camera data is used for retrieving the infrared face registry in the step 2 (all infrared face comparison presets are lower than a hit threshold value during retrieval) and the RGB camera cannot detect the face in the step 1, returning to the step 1 to restart a new face detection attempt, and if the RGB camera successfully detects the face in the step 1, entering the step 2 to further use the face characteristic value corresponding to the RGB camera data to retrieve the RGB face registry;
and 5, if the RGB face library retrieval processing in the step 4 fails to identify any person, the person cannot be effectively identified by the identification, and the step 1 is returned to restart a new face detection attempt. If the RGB face library retrieval processing in step 4 successfully matches the face features (higher than the preset hit threshold), an infrared face picture quality evaluation module is invoked to perform calculation, and when the quality of the current infrared face picture acquired simultaneously with the RGB camera is higher than the picture quality of the person in the infrared face registry, infrared face data corresponding to the person is added to the infrared face registry (this key processing completes the initial automatic registration of the infrared face registration data of the person originally having only RGB face registration data). And after the automatic registration processing of the infrared registration photo of the person is completed, the display module displays that the identification is successful.
Preferably, the usability statistic module is configured to perform the analysis based on recognition passing factors corresponding to the infrared registered photos of the person during daily use. In some embodiments, the analysis is performed based on the probability of recognition corresponding to the infrared registered photograph of the person, for example, 10 detections, 8 detections fail face authentication, and only 2 detections indicate that the recognition is relatively low in passing and needs to be replaced. This threshold may be preset or may be preset in a dynamic update manner.
The infrared photo quality evaluation module can adopt evaluation algorithms in the prior art, such as histogram-based information capacity detection method, energy spectrum-entropy detection method, and the like, and aims to perform quality evaluation on two compared infrared photos.
Therefore, the invention can automatically check whether the corresponding personnel has infrared face registration data after successfully identifying the personnel in the RGB face registry. If no infrared face registration data exists, automatically registering the infrared face data which currently meets the registration quality requirement after the infrared face imaging quality is automatically evaluated; and after the corresponding personnel of the infrared face registry is successfully identified, the availability evaluation of the infrared face registry photos of the personnel is automatically carried out. If the infrared face registration photo of the person needs to be optimized in evaluation, after the infrared face imaging quality is automatically evaluated, the infrared face registration data of the person is automatically updated by using the infrared face data which currently meets the registration quality requirement and has quality superior to that of the registered infrared face data of the person.
Therefore, the method for dynamically optimizing the influence of light in the face recognition can solve the problem of pre-registration of infrared face data and an automatic quality optimization mechanism of the infrared face registration data, form a high-availability infrared face registry, and provide a data base for the face recognition by using an infrared camera. The infrared camera is used for carrying out the whole-process face recognition processing, so that the face recognition processing is basically not influenced by external light conditions, and the face recognition technology and the environment robustness of the application of each field can be greatly improved.
In the traditional face recognition process, the registration of the infrared photo can only be directly registered at the equipment end, and the efficiency is low. The method of the invention can directly import the normal color photo registration and can be used after issuing. The infrared enrollment photographs are then automatically enrolled or optimized step by step during normal use.
Moreover, in the daily comparison process, the situation that the similarity between the registered photo and the user is always low can be met, and by adopting the automatic optimization updating method in the method, whether the photo needs to be optimized into a new infrared registered photo or not can be determined through statistical analysis in daily use.
Although the present invention has been described with reference to the preferred embodiments, it is not intended to be limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, the protection scope of the present invention should be determined by the appended claims.
Claims (10)
1. A method for dynamically optimizing light influence in face recognition is characterized by comprising the following steps:
step 1, simultaneously acquiring data of two cameras through a binocular camera with an RGB camera and an infrared camera, and extracting a face characteristic value after a face is detected;
step 2, preferentially using the infrared camera data to perform face detection processing, retrieving an infrared face registry to perform face recognition, entering the bottom library photo optimization of the infrared face registry if the face features in the infrared face registry are successfully matched, and performing face data re-acquisition or entering the face detection processing of the data of the RGB camera according to whether the data of the RGB camera detects the face if the matching is unsuccessful;
the method comprises the steps that a base photo of an infrared face registry is optimized, wherein the base photo optimization of the infrared face registry comprises the step of dynamically replacing an infrared face photo which is registered in the infrared face registry by using a currently obtained infrared face photo;
the face detection processing of the data of the RGB camera comprises the steps of responding to the face characteristics successfully matched into the RGB face registry, performing infrared photo quality detection, and adding and registering infrared face data corresponding to the person to the infrared face registry when the quality of the infrared face photos currently and simultaneously acquired by the RGB camera is higher than that of the person in the infrared face registry.
2. The method of claim 1, wherein in the step 2, the process of optimizing the base photo of the infrared face registry comprises:
calling an availability statistical module of the infrared face registry, detecting the registered infrared face photos, and returning a result of whether the photo optimization of the infrared face registry is needed:
if the registered infrared face photos in the infrared face registry do not need to be updated, the display module directly displays the successful recognition;
if the photo of the infrared face registry is required to be optimized, an infrared face photo quality evaluation module is called to detect, when the quality of the currently acquired infrared face photo is higher than that of the infrared face photo of the person in the infrared face registry, dynamic replacement is carried out, and the currently acquired infrared face photo is used for dynamically replacing the infrared face photo registered in the infrared face registry; and after the infrared registration photo is optimized, the display module displays that the identification is successful.
3. The method for dynamically optimizing the influence of light rays in face recognition according to claim 1, wherein in the step 2, whether a face is detected by the data of the RGB camera is judged, if not, the step 1 is returned to perform the re-acquisition of the face data, otherwise, the face detection processing of the data of the RGB camera is performed.
4. The method as claimed in claim 1, wherein the face detection processing of the RGB camera data comprises: and if the face features in the RGB face registry are not successfully matched, returning to the step 1 to carry out the re-acquisition of the face data.
5. The method for dynamically optimizing the influence of light rays in face recognition according to claim 1, wherein in the step 2, the infrared image quality detection comprises:
and (3) calling an infrared human face photo quality evaluation module for detection, and returning to the step (1) for re-acquisition of human face data if the quality of the infrared human face photos currently acquired simultaneously with the RGB camera is lower than that of the human face photos of the person in the infrared human face registry.
6. The method of claim 1, wherein the first automatic registration of the ir registered photo of the person is completed after the ir face data corresponding to the person is added and registered to the ir face registry, and the display module displays that the recognition is successful.
7. A system for dynamically optimizing light effects in face recognition, comprising:
the binocular camera arranged on the face recognition equipment comprises an RGB (red, green and blue) camera and an infrared camera, when a person passes through the binocular camera, data of the two cameras are simultaneously acquired through the RGB camera and the infrared camera, and after a face is detected, a face characteristic value is extracted;
at least one computing device, configured to preferentially use the infrared camera data to perform face detection processing, retrieve the infrared face registry to perform face recognition, enter the base photo optimization of the infrared face registry if the face features successfully matched in the infrared face registry, and perform face data re-acquisition or enter the face detection processing of the RGB camera data according to whether the data of the RGB camera detects a face if the matching is unsuccessful;
the method comprises the steps that a base photo of an infrared face registry is optimized, wherein the base photo optimization of the infrared face registry comprises the step of dynamically replacing an infrared face photo which is registered in the infrared face registry by using a currently obtained infrared face photo;
the face detection processing of the data of the RGB camera comprises the steps of responding to the face characteristics successfully matched into the RGB face registry, performing infrared photo quality detection, and adding and registering infrared face data corresponding to the person to the infrared face registry when the quality of the infrared face photos currently and simultaneously acquired by the RGB camera is higher than that of the person in the infrared face registry.
8. The system for dynamically optimizing ray effects in face recognition according to claim 7, wherein the at least one computing device is further configured to:
calling an availability statistical module of the infrared face registry, detecting the registered infrared face photos, and returning a result of whether the photo optimization of the infrared face registry is needed:
if the registered infrared face photos in the infrared face registry do not need to be updated, the display module directly displays the successful recognition;
if the photo of the infrared face registry is required to be optimized, an infrared face photo quality evaluation module is called to detect, when the quality of the currently acquired infrared face photo is higher than that of the infrared face photo of the person in the infrared face registry, dynamic replacement is carried out, and the currently acquired infrared face photo is used for dynamically replacing the infrared face photo registered in the infrared face registry; and after the infrared registration photo is optimized, the display module displays that the identification is successful.
9. The system for dynamically optimizing ray effects in face recognition according to claim 7, wherein the at least one computing device is further configured to:
and after the infrared face data corresponding to the registered person is added to the infrared face registry, the first automatic registration processing of the infrared registered photo of the person is completed, and the successful recognition is displayed.
10. The system for dynamically optimizing ray effects in face recognition according to any one of claims 7-9, wherein the binocular camera is disposed on the at least one computing device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010051299.9A CN110874588B (en) | 2020-01-17 | 2020-01-17 | Method and device for dynamically optimizing light influence in face recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010051299.9A CN110874588B (en) | 2020-01-17 | 2020-01-17 | Method and device for dynamically optimizing light influence in face recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110874588A true CN110874588A (en) | 2020-03-10 |
| CN110874588B CN110874588B (en) | 2020-04-14 |
Family
ID=69717613
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010051299.9A Active CN110874588B (en) | 2020-01-17 | 2020-01-17 | Method and device for dynamically optimizing light influence in face recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110874588B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112364842A (en) * | 2020-12-24 | 2021-02-12 | 杭州宇泛智能科技有限公司 | Double-shot face recognition method and device |
| CN113963392A (en) * | 2020-07-03 | 2022-01-21 | 北京君正集成电路股份有限公司 | Face recognition method based on dynamic adjustment threshold |
| US11443550B2 (en) | 2020-06-05 | 2022-09-13 | Jilin Qs Spectrum Data Technology Co. Ltd | Face recognition monitoring system based on spectrum and multi-band fusion and recognition method using same |
| CN113963392B (en) * | 2020-07-03 | 2024-05-03 | 北京君正集成电路股份有限公司 | Face recognition method based on dynamic adjustment threshold |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101964056A (en) * | 2010-10-26 | 2011-02-02 | 徐勇 | Bimodal face authentication method with living body detection function and system |
| CN106204815A (en) * | 2016-06-23 | 2016-12-07 | 江西洪都航空工业集团有限责任公司 | A kind of gate control system based on human face detection and recognition |
| CN108171834A (en) * | 2017-12-25 | 2018-06-15 | 深圳禾思众成科技有限公司 | A kind of intelligent access control system |
| CN108875476A (en) * | 2017-08-03 | 2018-11-23 | 北京旷视科技有限公司 | Automatic near-infrared face registration and recognition methods, device and system and storage medium |
| CN108875338A (en) * | 2018-05-04 | 2018-11-23 | 北京旷视科技有限公司 | unlocking method, device and system and storage medium |
| CN109886222A (en) * | 2019-02-26 | 2019-06-14 | 北京市商汤科技开发有限公司 | Face identification method, neural network training method, device and electronic equipment |
-
2020
- 2020-01-17 CN CN202010051299.9A patent/CN110874588B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101964056A (en) * | 2010-10-26 | 2011-02-02 | 徐勇 | Bimodal face authentication method with living body detection function and system |
| CN106204815A (en) * | 2016-06-23 | 2016-12-07 | 江西洪都航空工业集团有限责任公司 | A kind of gate control system based on human face detection and recognition |
| CN108875476A (en) * | 2017-08-03 | 2018-11-23 | 北京旷视科技有限公司 | Automatic near-infrared face registration and recognition methods, device and system and storage medium |
| CN108171834A (en) * | 2017-12-25 | 2018-06-15 | 深圳禾思众成科技有限公司 | A kind of intelligent access control system |
| CN108875338A (en) * | 2018-05-04 | 2018-11-23 | 北京旷视科技有限公司 | unlocking method, device and system and storage medium |
| CN109886222A (en) * | 2019-02-26 | 2019-06-14 | 北京市商汤科技开发有限公司 | Face identification method, neural network training method, device and electronic equipment |
Non-Patent Citations (1)
| Title |
|---|
| ANDREAS ELLMAUTHALER 等: "A visible-light and infrared video database for performance evaluation of video/image fusion methods", 《MULTIDIMENSIONAL SYSTEMS AND SIGNAL PROCESSING》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11443550B2 (en) | 2020-06-05 | 2022-09-13 | Jilin Qs Spectrum Data Technology Co. Ltd | Face recognition monitoring system based on spectrum and multi-band fusion and recognition method using same |
| CN113963392A (en) * | 2020-07-03 | 2022-01-21 | 北京君正集成电路股份有限公司 | Face recognition method based on dynamic adjustment threshold |
| CN113963392B (en) * | 2020-07-03 | 2024-05-03 | 北京君正集成电路股份有限公司 | Face recognition method based on dynamic adjustment threshold |
| CN112364842A (en) * | 2020-12-24 | 2021-02-12 | 杭州宇泛智能科技有限公司 | Double-shot face recognition method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110874588B (en) | 2020-04-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2019033572A1 (en) | Method for detecting whether face is blocked, device and storage medium | |
| WO2019109526A1 (en) | Method and device for age recognition of face image, storage medium | |
| JP4505362B2 (en) | Red-eye detection apparatus and method, and program | |
| CN106462766B (en) | Image capture parameters adjustment is carried out in preview mode | |
| JP3753722B2 (en) | Extraction method of tooth region from tooth image and identification method and apparatus using tooth image | |
| JP4725377B2 (en) | Face image registration device, face image registration method, face image registration program, and recording medium | |
| WO2018233038A1 (en) | Deep learning-based method, apparatus and device for recognizing license plate, and storage medium | |
| EP2806373A2 (en) | Image processing system and method of improving human face recognition | |
| WO2019033574A1 (en) | Electronic device, dynamic video face recognition method and system, and storage medium | |
| US20060110014A1 (en) | Expression invariant face recognition | |
| CN110874588B (en) | Method and device for dynamically optimizing light influence in face recognition | |
| US20180276487A1 (en) | Method, system, and computer-readable recording medium for long-distance person identification | |
| CN102859554B (en) | Collating device | |
| CN101887518B (en) | Human detecting device and method | |
| CN104185849A (en) | Method and device for notification of facial recognition environment, and computer-readable recording medium for executing method | |
| US20110141257A1 (en) | Apparatus and method for registering plurality of facial images for face recognition | |
| CN106295672A (en) | A kind of face identification method and device | |
| CN106228168A (en) | The reflective detection method of card image and device | |
| US11678180B2 (en) | Iris recognition workflow | |
| CN111523383B (en) | Non-perception face recognition system and method based on pedestrian ReID | |
| CN110532746B (en) | Face checking method, device, server and readable storage medium | |
| US20200364479A1 (en) | Face recognition system, method for establishing data of face recognition, and face recognizing method thereof | |
| Su et al. | Robust video fingerprinting based on visual attention regions | |
| CN112434638A (en) | Facial expression recognition, classification and analysis system | |
| JP2014128003A (en) | Image recognition device, and control method and program thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder |
Address after: No.568 longmian Avenue, gaoxinyuan, Jiangning District, Nanjing City, Jiangsu Province, 211000 Patentee after: Xiaoshi Technology (Jiangsu) Co.,Ltd. Address before: No.568 longmian Avenue, gaoxinyuan, Jiangning District, Nanjing City, Jiangsu Province, 211000 Patentee before: NANJING ZHENSHI INTELLIGENT TECHNOLOGY Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |