WO2017054276A1 - 一种生物特征身份识别方法及装置 - Google Patents
一种生物特征身份识别方法及装置 Download PDFInfo
- Publication number
- WO2017054276A1 WO2017054276A1 PCT/CN2015/093379 CN2015093379W WO2017054276A1 WO 2017054276 A1 WO2017054276 A1 WO 2017054276A1 CN 2015093379 W CN2015093379 W CN 2015093379W WO 2017054276 A1 WO2017054276 A1 WO 2017054276A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- biometric
- image
- matching
- regions
- biometric image
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000012795 verification Methods 0.000 title abstract 5
- 238000000605 extraction Methods 0.000 claims description 30
- 230000000903 blocking effect Effects 0.000 claims description 9
- 238000005286 illumination Methods 0.000 abstract description 10
- 230000007547 defect Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 6
- 238000010606 normalization Methods 0.000 description 5
- 210000001747 pupil Anatomy 0.000 description 5
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 238000003384 imaging method Methods 0.000 description 3
- 238000005192 partition Methods 0.000 description 2
- 230000005489 elastic deformation Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/32—User authentication using biometric data, e.g. fingerprints, iris scans or voiceprints
-
- 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
- G06V40/12—Fingerprints or palmprints
- G06V40/13—Sensors therefor
-
- 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
-
- 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
- G06V40/18—Eye characteristics, e.g. of the iris
-
- 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
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/193—Preprocessing; Feature extraction
-
- 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
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/197—Matching; Classification
Definitions
- the present invention belongs to the field of feature recognition technologies, and in particular, to a biometric identity identification method and apparatus.
- iris recognition Due to the stability, non-invasiveness and non-changeability of the iris itself, iris recognition has become a research hotspot in the field of biometric identity identification.
- the main processes of iris recognition include: iris localization, iris image acquisition, iris image normalization, iris image enhancement and iris image feature extraction, and then the extracted features are matched to obtain the iris image corresponding user ID.
- the iris positioning mainly identifies the inner and outer boundaries of the eye through the recognition algorithm, and locates the iris through the inner and outer boundaries.
- the inner boundary mainly refers to the boundary of the pupil. Since the brightness of the pupil and the eye are different, the pupil is usually used by binarization. Separation in the eye, the outer boundary refers to the boundary of the circle concentric with the pupil to the upper and lower boundaries of the eye.
- the normalization of the iris image is to uniformly adjust the size and proportion of the iris image to a uniform calibration. This is because the pupil zoom caused by the illumination change and the distance between the lens and the human eye are different, which will cause the iris to have different sizes in the human eye.
- This elastic deformation will affect the extraction of features from the iris image, so the size and proportion of the iris image must be uniformly adjusted to a uniform calibration.
- the current common normalization method is that the iris image under the "two-dimensional coordinate system" changes to the iris image under the "polar coordinate system". And because of the uneven illumination, the brightness of each pixel in the iris image is different, which increases the difficulty of extracting features from the iris image, so the iris image can be enhanced by the equalization of the luminance histogram, so that different regions are The brightness of each pixel becomes similar, simplifying the difficulty of feature extraction.
- Corresponding feature extraction is performed on the entire iris image in the feature extraction process. For example, extracting the position, size and shape of the texture from an iris image as the feature of iris recognition, that is, the existing feature extraction is performed after normalizing and enhancing the iris image. These three features are extracted from the entire iris image. If the image information of the iris image is incomplete due to factors such as illumination, angle and distance during the process of image acquisition, the complete iris feature cannot be extracted. In this case, the iris image is repeatedly acquired for recognition.
- an object of the present invention is to provide a biometric identification method and apparatus for reducing the number of repeated recognitions.
- the technical solutions are as follows:
- the invention provides a biometric identification method, the method comprising:
- the first biometric image is segmented to obtain a plurality of first biometric regions
- Feature extraction is performed on each of the first biometric regions to obtain feature data of each of the first biometric regions;
- Feature matching is performed on feature data of each first biometric area, wherein the feature matching process is: matching feature data of the first biometric area of the i-th block with feature data of the second biometric area of the i-th block stored in advance Obtaining a matching result of the first biometric region of the i-th block, 1 ⁇ i ⁇ n, n is the total number of the first bio-feature regions obtained, and the pre-stored i-th block second bio-feature region is a pre-stored legal An area obtained by segmenting the second biometric image of the user;
- the first biometric image is segmented to obtain a plurality of first biometric regions, including:
- the first biometric image After obtaining the first biometric image, the first biometric image is segmented according to a preset number of blocks to obtain a plurality of first biometric regions.
- performing feature matching on the feature data of each of the first biometric regions comprises: matching feature data of the plurality of first biometric regions in parallel.
- the characterizing the feature data of each piece of the first biometric area comprises: matching the feature data of each piece of the first biometric area in sequence.
- the method further includes: after matching the preset number of first biometric regions, determining whether the matching degree of the first biometric image obtained based on the matching result of the preset number of first biometric regions indicates The user identity recognition result corresponding to the first biometric image is legal;
- the invention also provides a biometric identification device, the device comprising:
- a blocking unit configured to: after obtaining the first biometric image, segment the first biometric image to obtain a plurality of first biometric regions;
- An extracting unit configured to perform feature extraction on each of the first biometric regions, to obtain feature data of each first biometric region
- a matching unit configured to perform feature matching on the feature data of each of the first biometric regions, wherein the feature matching process is: selecting feature data of the first biometric region of the i-th block and pre-stored second biometric region of the i-th block The feature data is matched to obtain a matching result of the first biometric region of the i-th block, 1 ⁇ i ⁇ n, n is the total number of the first bio-feature regions obtained, and the pre-stored i-th block second bio-feature region is An area obtained by dividing a second biometric image of a legal user stored in advance;
- a calculating unit configured to obtain a matching degree of the first biometric image based on a matching result of the first biometric region, wherein a matching degree of the first biometric image is used to indicate the first biometric image a match with respect to the second biometric image;
- an identifying unit configured to obtain a user identity recognition result corresponding to the first biometric image based on a matching degree of the first biometric image.
- the blocking unit is configured to: after obtaining the first biometric image, segment the first biometric image according to a preset number of blocks to obtain a plurality of first biometric regions.
- the extracting unit is configured to: perform matching on the feature data of the plurality of first biometric regions in parallel.
- the extracting unit is configured to: perform matching on feature data of each piece of the first biometric area in sequence.
- the device further includes: a determining unit, configured to determine, according to the matching of the preset number of first biometric regions, the first biometric image obtained by the matching result of the preset number of first biometric regions Whether the matching degree indicates that the user identification result corresponding to the first biometric image is legal, and if yes, triggering the extracting unit to terminate matching of the remaining first biometric regions.
- a determining unit configured to determine, according to the matching of the preset number of first biometric regions, the first biometric image obtained by the matching result of the preset number of first biometric regions Whether the matching degree indicates that the user identification result corresponding to the first biometric image is legal, and if yes, triggering the extracting unit to terminate matching of the remaining first biometric regions.
- the first biometric image is segmented to obtain a plurality of first biometric regions, and then feature extraction is performed on each of the first biometric regions to obtain each And acquiring the feature of the first biometric image based on the matching result of the first biometric region, and obtaining the matching degree of the first biometric image according to the matching degree of the first biometric image, and obtaining the corresponding corresponding to the first biometric image.
- User identification result That is to say, the invention can obtain the user identification result corresponding to the first biometric image based on the proportion of the legal matching of the first biometric region, thereby reducing the quality requirement of the first biometric image to a certain extent, thereby reducing the shooting.
- Angles, illumination, etc. result in the probability of repeated recognition resulting from the incompleteness of the first biometric image.
- FIG. 1 is a flowchart of a biometric identity identification method according to an embodiment of the present invention
- FIG. 2 is a schematic diagram of partitioning a first biometric image according to an embodiment of the present invention
- FIG. 3 is a schematic diagram of a first biometric area matching process according to an embodiment of the present invention.
- FIG. 4 is a schematic diagram of an iris image matching process and a matching result according to an embodiment of the present invention.
- FIG. 5 is a schematic structural diagram of a biometric identity recognition apparatus according to an embodiment of the present invention.
- FIG. 6 is another schematic structural diagram of a biometric identity recognition apparatus according to an embodiment of the present invention.
- FIG. 1 is a flowchart of a biometric identification method provided by an embodiment of the present invention.
- the feature extraction may be performed in a manner different from the method of extracting feature data directly from the entire first biometric image. Includes the following steps:
- the first biometric image is segmented to obtain a plurality of first biometric regions. That is to say, each of the first biometric regions is a partial region of the first biometric image, and each of the first biometric regions is obtained by segmenting the first biometric image.
- the first biometric image is a basis for identifying a user identity, and may be any one of a face image, an iris image, and a fingerprint image.
- the face image and the iris image can be acquired by the camera device in the electronic device, and the fingerprint image can be acquired by the fingerprint collection device in the electronic device.
- the first biometric image is an iris image or a fingerprint image
- the first biometric image may be directly segmented; in the case where the first biometric image is a face image, the face image needs to be first recognized.
- the area where the feature object is located such as the area where the human eye, the mouth, and the nose are located, and then the area where the feature object is located is divided to obtain a plurality of first biometric areas.
- the first biometric image may be segmented according to a preset number of segments, that is, the first biometric image is divided into a first biometric image of a preset number of segments, As shown in FIG. 2, wherein the left side of FIG. 2 is a first biometric image, the first biometric image is divided into n first biometric regions shown on the right side of FIG. 2, where n is a preset number of partitions.
- the preset number of blocks may be a fixed value or may be based on The accuracy is set.
- the recognition accuracy is high, the value of the preset number of blocks can be increased.
- the recognition accuracy is low, the value of the preset number of blocks can be reduced.
- the preset number of blocks may also be related to the quality of the first biometric image, that is, the value of the preset number of blocks may be increased when the quality of the first biometric image is high; when the quality of the first biometric image is When the value is lower, the value of the preset number of blocks can be reduced, wherein the quality of the first biometric image is related to the configuration parameter of the device for acquiring the first biometric image, and the imaging device is taken as an example.
- the better the definition of the first biometric image is, the higher the quality of the first biometric image is; when the resolution of the camera is lower, the sharpness of the obtained first biometric image is Lower, indicating that the quality of the first biometric image is lower.
- the block extraction method can enable the electronic device to perform feature extraction on the plurality of first biometric regions in parallel, thereby Speed up the extraction.
- 103 Perform feature matching on feature data of each piece of the first biometric area, wherein the feature matching process is: selecting feature data of the first biometric area of the i-th block and characteristic data of the second biometric area of the i-th block pre-stored Matching is performed to obtain a matching result of the first biometric region of the i-th block, 1 ⁇ i ⁇ n, n is the total number of the first biometric regions obtained, and the pre-stored second biometric region of the i-th block is pre-stored legally.
- the user's second biometric image is segmented to obtain an area.
- the number of second biometric regions of the second biometric image and the first biometric region of the first biometric image is the same, so that the first biometric region and the second biometric region can be matched one-to-one.
- the process of one-to-one matching between the first biometric region and the second biometric region is as shown in FIG.
- 105 Obtain a user identity recognition result corresponding to the first biometric image based on the matching degree of the first biometric image.
- the matching degree of the first biometric image is used to indicate a matching situation of the first biometric image with respect to the second biometric image, if the matching degree of the first biometric image indicates that the first biometric image matches the second biometric image , indicating that the user identity corresponding to the first biometric image is a legal user identity; if the matching degree of the first biometric image indicates that the first biometric image does not match the second biometric image, indicating that the first biometric image corresponds to The user identity is an illegal user.
- the biometric identification method after obtaining the first biometric image, performs segmentation on the first biometric image to obtain a plurality of first biometric regions, and then first for each block. Feature extraction is performed on the biometric region, and feature data of each first biometric region is obtained, and then the matching degree of the first biometric image is obtained based on the matching result of the first biometric region, and based on the matching degree of the first biometric image.
- the invention can obtain the user identity recognition result corresponding to the first biometric image based on the proportion of legal matching of the first biometric region, which is reduced to some extent The quality requirement of the first biometric image, thereby reducing the probability of repeated recognition caused by inaccurate acquisition of the first biometric image due to shooting angle, illumination, and the like.
- the matching degree of the first biometric image can be obtained by normalizing the matching result of each first biometric region, and the normalization processing process is as follows:
- the determination threshold of the first biometric area is set as ⁇ , wherein the value of ⁇ may be determined according to the security level requirement of the electronic device. For example, when the security level requires a lower application, ⁇ may take a smaller value, such as 60%, on the contrary, for applications with higher security levels, ⁇ can take 80% to 90%.
- the matching of each first biometric area is represented by Q i , and the values are as follows:
- the value "1" indicates that the first biometric region of the i-th block matches the second biometric region of the i-th block; the value "0" indicates that the first biometric region of the i-th block and the second bio-characteristic region of the i-th block are not match.
- n is the total number of blocks and ⁇ is the degree of matching of the first biometric image. That is to say, in the embodiment of the present invention, the matching degree of the first biometric image is related to the number of matched first biometric regions, and the implementation of the present invention is performed when the first biometric image is incomplete due to the shooting angle, illumination, or the like. For example, the number of first biometric regions matching the second biometric region can be increased by dividing more first biometric regions, thereby improving the primary matching manner based on the entire first biometric image. Identify accuracy.
- the first biometric image is assumed to be an iris image
- the feature data of the two regions and the feature data of the second biometric image are:
- Lower half area feature data ⁇ 47, 35, 47, 38, 69, 85, 25, 35, 99, 52 ⁇ ;
- the upper half region feature sequence of the second biometric image ⁇ 50, 78, 32, 41, 65, 85, 49, 55, 21, 11 ⁇ ;
- the lower half region feature sequence of the second biometric image ⁇ 33, 45, 47, 77, 39, 41, 25, 85, 22, 58 ⁇ ;
- the matching process of the two areas is shown in Figure 4, in which the data of the bold part indicates mismatch, the data of the normal font part matches, and the matching degree of the upper part is 80%, because the iris image is affected by the illumination and the shooting angle.
- the image quality collected in the lower half is poor, so that the features extracted in the lower half are not accurate enough, and finally the matching degree in the lower half is 20%. If it is directly from the entire iris The image is feature extracted and matched with a matching degree of up to 50%.
- the biometric identification method provided by the embodiment of the present invention may further match the feature data of each first biometric region in a serial manner, that is, for the first biometric region of the first block to the first biometric region of the nth block. Said to sequentially match the feature data of each of the first biometric regions, so that after matching the preset number of first biometric regions, determining the matching result based on the matching number of the first biometric regions of the preset number Whether the matching degree of the biometric image indicates that the user identification result corresponding to the first biometric image is legal, and if so, the matching of the remaining first biometric regions is terminated; if not, the remaining first biometric region may be In the process of performing the matching in sequence, determining whether the matching degree of the first biometric image obtained after each matching indicates that the user identification result corresponding to the first biometric image is legal, and if yes, terminating the remaining first biometric region Match, if otherwise continue to match and judge.
- the first biometric area may be matched first, and then the first biometric image based on the part of the first biometric area is obtained.
- the matching degree if the matching degree of the first biometric image obtained at this time indicates that the user identification result corresponding to the first biometric image is legal, the remaining first biometric regions may not be matched, thereby reducing the matching.
- the number of first biometric regions can save system resources.
- the preset number may be greater than or equal to n/2, and the specific value is not limited in the embodiment of the present invention.
- the matching degree of the first biometric image provided by the embodiment of the present invention is related to the number of matched first biometric regions, and if only one of the one hundred first biometric regions matches, even if it The matching degree reaches 100%, and the user identity is also determined to be an illegal user identity, so as to eliminate misidentification caused by inaccurate feature data extracted. If the first biometric is required to match If there are too many areas, the meaning of blocking will be lost. Therefore, the value of ⁇ should not be higher in the setting process. For example, in the application with higher level requirements, the maximum value of ⁇ can be 70%, and the application with low security level is not high. A slightly smaller ratio such as 30% can be selected.
- the embodiment of the present invention further provides a biometric identity identification device.
- the schematic diagram of the structure is as shown in FIG. 5, and may include: a blocking unit 11, an extracting unit 12, a matching unit 13, and a computing unit. 14 and identification unit 15.
- the blocking unit 11 is configured to, after obtaining the first biometric image, segment the first biometric image to obtain a plurality of first biometric regions. That is to say, each of the first biometric regions is a partial region of the first biometric image, and each of the first biometric regions is obtained by segmenting the first biometric image.
- the first biometric image is a basis for identifying a user identity, and may be any one of a face image, an iris image, and a fingerprint image.
- the face image and the iris image can be acquired by the camera device in the electronic device, and the fingerprint image can be acquired by the fingerprint collection device in the electronic device.
- the first biometric image is an iris image or a fingerprint image
- the first biometric image may be directly segmented; in the case where the first biometric image is a face image, the face image needs to be first recognized.
- the area where the feature object is located such as the area where the human eye, the mouth, and the nose are located, and then the area where the feature object is located is divided to obtain a plurality of first biometric areas.
- the blocking unit 11 may block the first biometric image according to the preset number of blocks, that is, divide the first biometric image into the first number of preset segments.
- the biometric image is as shown in FIG. 2, wherein the left side of FIG. 2 is a first biometric image, and the first biometric image is divided into n first biometric regions shown on the right side of FIG. 2, where n is a preset score. The number of blocks.
- the preset number of blocks may be a fixed value or may be set according to the recognition accuracy.
- the recognition accuracy is high, the value of the preset number of blocks may be increased;
- the accuracy is low, the value of the preset number of blocks can be reduced.
- the above preset number of blocks is also It may be related to the quality of the first biometric image, that is, the value of the preset number of blocks may be increased when the quality of the first biometric image is high; and the preset may be lowered when the quality of the first biometric image is low.
- the value of the number of partitions, wherein the quality of the first biometric image is related to the configuration parameter of the device for acquiring the first biometric image, and taking the imaging device as an example, when the resolution of the imaging device is high, the first biological creature is obtained.
- the better the definition of the feature image the higher the quality of the first biometric image; when the resolution of the camera is lower, the sharpness of the obtained first biometric image is reduced, indicating the first biometric image The quality is lower.
- the extracting unit 12 is configured to perform feature extraction on each of the first biometric regions to obtain feature data of each of the first biometric regions, that is, the embodiment of the present invention may perform segmentation to extract feature data in the first biometric image.
- the block extraction method can be used to make the extraction unit 12 parallel to the plurality of first biometric regions. Feature extraction to speed up extraction.
- the matching unit 13 is configured to perform feature matching on the feature data of each of the first biometric regions, wherein the feature matching process is: selecting feature data of the first biometric region of the i-th block and pre-stored second biometric features of the i-th block The feature data of the region is matched to obtain a matching result of the first biometric region of the i-th block, 1 ⁇ i ⁇ n, n is the total number of the first bio-feature regions obtained, and the pre-stored second bio-feature region of the i-th block is A region obtained by segmenting the second biometric image of the legal user stored in advance.
- the number of second biometric regions of the second biometric image and the first biometric region of the first biometric image is the same, so that the first biometric region and the second biometric region can be matched one-to-one.
- the process of one-to-one matching between the first biometric region and the second biometric region is as shown in FIG.
- the calculating unit 14 is configured to obtain a matching degree of the first biometric image based on a matching result of the first biometric region.
- the identifying unit 15 is configured to obtain a user identity recognition result corresponding to the first biometric image based on the matching degree of the first biometric image.
- the matching degree of the first biometric image is used to indicate a matching situation of the first biometric image with respect to the second biometric image, if the matching degree of the first biometric image indicates that the first biometric image matches the second biometric image , indicating that the user identity corresponding to the first biometric image is a legal user identity; if the matching degree of the first biometric image indicates that the first biometric image does not match the second biometric image, indicating that the first biometric image corresponds to The user identity is an illegal user.
- the biometric identification device after obtaining the first biometric image, performs segmentation on the first biometric image to obtain a plurality of first biometric regions, and then first for each block. Feature extraction is performed on the biometric region, and feature data of each first biometric region is obtained, and then the matching degree of the first biometric image is obtained based on the matching result of the first biometric region, and based on the matching degree of the first biometric image. Obtaining a user identification result corresponding to the first biometric image.
- the invention can obtain the user identification result corresponding to the first biometric image based on the proportion of the legal matching of the first biometric region, thereby reducing the quality requirement of the first biometric image to a certain extent, thereby reducing the shooting.
- Angles, illumination, etc. result in the probability of repeated recognition resulting from the incompleteness of the first biometric image.
- the matching degree of the first biometric image can be obtained by normalizing the matching result of each first biometric region, and the normalization processing process is as follows:
- the determination threshold of the first biometric area is set as ⁇ , wherein the value of ⁇ may be determined according to the security level requirement of the electronic device. For example, when the security level requires a lower application, ⁇ may take a smaller value, such as 60%, on the contrary, for applications with higher security levels, ⁇ can take 80% to 90%.
- the matching of each first biometric area is represented by Q i , and the values are as follows:
- the value "1" indicates that the first biometric region of the i-th block matches the second biometric region of the i-th block; the value "0" indicates that the first biometric region of the i-th block and the second bio-characteristic region of the i-th block are not match.
- the matching degree of the first biometric image is related to the number of matched first biometric regions, and the implementation of the present invention is performed when the first biometric image is incomplete due to the shooting angle, illumination, or the like.
- the number of first biometric regions matching the second biometric region can be increased by dividing more first biometric regions, thereby improving the primary matching manner based on the entire first biometric image. Identify accuracy.
- the biometric identification device provided by the embodiment of the present invention has other structures, as shown in FIG. 6, wherein the feature of the first biometric region is extracted serially by the extracting unit 12 in the biometric identification device shown in FIG. The data is matched, that is, for the first biometric region to the nth first biometric region of the first block, the extracting unit 12 sequentially matches the feature data of each of the first biometric regions.
- the determining unit 16 is configured to determine, after matching the preset number of first biometric regions, whether the matching degree of the first biometric image obtained by the matching result of the preset number of first biometric regions indicates the first creature The user identification result corresponding to the feature image is legal. If yes, the trigger extraction unit 12 terminates the matching of the remaining first biometric regions; if otherwise, the trigger extraction unit 12 sequentially matches the remaining first biometric regions, and In the process of sequentially matching the remaining first biometric regions by the extracting unit, the determining unit 16 still needs to determine whether the matching degree of the first biometric image obtained after each matching indicates the user identity corresponding to the first biometric image. The result is legal, and if so, the trigger extraction unit 12 terminates the matching of the remaining first biometric regions, if otherwise the trigger extraction unit 12 continues to match and determine.
- the first biometric area may be matched first, and then the matching degree of the first biometric image based on the part of the first biometric area is obtained, and if the matching degree of the first biometric image obtained at this time indicates the first If the user identification result corresponding to the biometric image is legal, the remaining first biometric regions may not be matched, thereby reducing the number of matching first biometric regions, thereby saving system resources.
- the preset number may be greater than or equal to n/2, and the specific value is not limited in the embodiment of the present invention.
- the matching degree of the first biometric image provided by the embodiment of the present invention is related to the number of matched first biometric regions, and if only one of the one hundred first biometric regions matches, even if it The matching degree reaches 100%, and the user identity is also determined to be an illegal user identity, so as to eliminate misidentification caused by inaccurate feature data extracted. If there is too much first biometric area to be matched, the meaning of blocking will be lost. Therefore, the value of ⁇ should not be higher in the setting process. For example, the maximum value of ⁇ can be 70% in applications with higher level requirements. A slightly smaller ratio, such as 30%, can be selected for applications where the security level is not critical.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Ophthalmology & Optometry (AREA)
- Computer Security & Cryptography (AREA)
- Computer Hardware Design (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Collating Specific Patterns (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
本发明提供一种生物特征身份识别方法及装置，在获得第一生物特征图像后，对第一生物特征图像进行分块得到多块第一生物特征区域，再对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据，然后基于第一生物特征区域的匹配结果，得到第一生物特征图像的匹配度，并基于第一生物特征图像的匹配度，得到第一生物特征图像对应的用户身份识别结果。也就是说本发明可以基于第一生物特征区域的合法匹配的比例来得到第一生物特征图像对应的用户身份识别结果，在一定程度上降低对第一生物特征图像的质量要求，进而降低因拍摄角度、光照等导致采集第一生物特征图像不完整导致的重复识别的概率。
Description
本发明属于特征识别技术领域，更具体的说，尤其涉及一种生物特征身份识别方法及装置。
由于虹膜自身的稳定性、非侵犯性、不可更改性等优点，虹膜识别已经成为生物特征身份鉴别领域中的研究热点。现有虹膜识别的主要过程包括：虹膜定位，虹膜图像的采集，虹膜图像的归一化，虹膜图像的增强和虹膜图像中的特征提取，然后将提取到的特征进行匹配以得到虹膜图像对应的用户身份。
其中，虹膜定位主要是通过识别算法来识别眼睛内外边界，通过内外边界定位虹膜，内边界主要是指瞳孔的边界，由于瞳孔与眼睛的亮度区别较大，通常采用二值化即可将瞳孔从眼睛中分离，外边界则是指与瞳孔同心的到眼睛上下边界的圆的边界。虹膜图像的归一化则是将虹膜图像的大小和比例统一调整到统一的校准，这是因为光照变化引起的瞳孔缩放以及镜头与人眼的距离不同，会造成虹膜在人眼图像的大小不同，这种弹性形变将会影响从虹膜图像中提取特征，因此必须将虹膜图像的大小和比例统一调整到统一的校准。目前常见的归一化方法是“二维坐标系”下的虹膜图像变化为“极坐标系”下的虹膜图像。并且由于光照的不均匀，导致虹膜图像中各个像素的亮度不一，这就增加从虹膜图像中提取特征的难度，所以可以通过亮度直方图的均衡化对虹膜图像进行增强处理，使得不同区域中各个像素的亮度变得相近，简化特征提取的难度。
相应的在特征提取过程中是以整幅虹膜图像为对象进行特征提取。比如从一幅虹膜图像中提取纹理的位置、大小、形状三个指标作为虹膜识别的特征，也就是说现有特征提取是在对虹膜图像进行归一化和增强处理后，直接
从整幅虹膜图像中提取这三个特征。若是采集图像的过程中受到光照、角度、距离等因素影响导致采集到的虹膜图像信息不完整，使得无法提取到完整的虹膜特征，在此种情况下就会重复获取虹膜图形进行识别。
发明内容
有鉴于此，本发明的目的在于提供一种生物特征身份识别方法及装置，用于降低重复识别的次数。技术方案如下：
本发明提供一种生物特征身份识别方法，所述方法包括：
在获得第一生物特征图像后，对所述第一生物特征图像进行分块，得到多块第一生物特征区域；
对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据；
对每块第一生物特征区域的特征数据进行特征匹配，其中特征匹配过程是：将第i块第一生物特征区域的特征数据与预先存储的第i块第二生物特征区域的特征数据进行匹配，得到第i块第一生物特征区域的匹配结果，1≤i≤n，n为得到的第一生物特征区域总数，所述预先存储的第i块第二生物特征区域是对预先存储的合法用户的第二生物特征图像进行分块得到的一个区域；
基于所述第一生物特征区域的匹配结果，得到所述第一生物特征图像的匹配度，其中所述第一生物特征图像的匹配度用于指示所述第一生物特征图像相对于所述第二生物特征图像的匹配情况；
基于所述第一生物特征图像的匹配度，得到所述第一生物特征图像对应的用户身份识别结果。
优选的，所述在获得第一生物特征图像后，对所述第一生物特征图像进行分块，得到多块第一生物特征区域，包括：
在获得第一生物特征图像后，按照预设分块数量对所述第一生物特征图像进行分块，得到多块第一生物特征区域。
优选的，所述对每块第一生物特征区域的特征数据进行特征匹配，包括：并行对多块第一生物特征区域的特征数据进行匹配。
优选的，所述对每块第一生物特征区域的特征数据进行特征匹配，包括：依次对每块第一生物特征区域的特征数据进行匹配。
优选的，所述方法还包括：在对预设数量的第一生物特征区域进行匹配后，判断基于预设数量的第一生物特征区域的匹配结果得到的第一生物特征图像的匹配度是否指示所述第一生物特征图像对应的用户身份识别结果合法；
如果是，则终止对剩余的第一生物特征区域的匹配。
本发明还提供一种生物特征身份识别装置，所述装置包括：
分块单元，用于在获得第一生物特征图像后，对所述第一生物特征图像进行分块，得到多块第一生物特征区域；
提取单元，用于对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据；
匹配单元，用于对每块第一生物特征区域的特征数据进行特征匹配，其中特征匹配过程是：将第i块第一生物特征区域的特征数据与预先存储的第i块第二生物特征区域的特征数据进行匹配，得到第i块第一生物特征区域的匹配结果，1≤i≤n，n为得到的第一生物特征区域总数，所述预先存储的第i块第二生物特征区域是对预先存储的合法用户的第二生物特征图像进行分块得到的一个区域；
计算单元，用于基于所述第一生物特征区域的匹配结果，得到所述第一生物特征图像的匹配度，其中所述第一生物特征图像的匹配度用于指示所述第一生物特征图像相对于所述第二生物特征图像的匹配情况；
识别单元，用于基于所述第一生物特征图像的匹配度，得到所述第一生物特征图像对应的用户身份识别结果。
优选的，所述分块单元用于：在获得第一生物特征图像后，按照预设分块数量对所述第一生物特征图像进行分块，得到多块第一生物特征区域。
优选的，所述提取单元用于：并行对多块第一生物特征区域的特征数据进行匹配。
优选的，所述提取单元用于：依次对每块第一生物特征区域的特征数据进行匹配。
优选的，所述装置还包括：判断单元，用于在对预设数量的第一生物特征区域进行匹配后，判断基于预设数量的第一生物特征区域的匹配结果得到的第一生物特征图像的匹配度是否指示所述第一生物特征图像对应的用户身份识别结果合法，如果是，则触发所述提取单元终止对剩余的第一生物特征区域的匹配。
与现有技术相比，本发明提供的上述技术方案具有如下优点：
本发明提供的上述技术方案，在获得第一生物特征图像后，对第一生物特征图像进行分块得到多块第一生物特征区域，再对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据，然后基于第一生物特征区域的匹配结果，得到第一生物特征图像的匹配度，并基于第一生物特征图像的匹配度，得到第一生物特征图像对应的用户身份识别结果。也就是说本发明可以基于第一生物特征区域的合法匹配的比例来得到第一生物特征图像对应的用户身份识别结果，在一定程度上降低对第一生物特征图像的质量要求，进而降低因拍摄角度、光照等导致采集第一生物特征图像不完整导致的重复识别的概率。
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本发明实施例提供的生物特征身份识别方法的流程图；
图2是本发明实施例提供的对第一生物特征图像分块的示意图；
图3是本发明实施例提供的第一生物特征区域匹配过程的示意图；
图4是本发明实施例提供的虹膜图像匹配过程和匹配结果的示意图；
图5是本发明实施例提供的生物特征身份识别装置的一种结构示意图；
图6是本发明实施例提供的生物特征身份识别装置的另一种结构示意图。
为使本发明实施例的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
请参阅图1，其示出了本发明实施例提供的生物特征身份识别方法的流程图，可以采用与现有直接从整幅第一生物特征图像中提取特征数据不同的方式进行特征提取，可以包括以下步骤：
101：在获得第一生物特征图像后，对第一生物特征图像进行分块，得到多块第一生物特征区域。也就是说每块第一生物特征区域是第一生物特征图像的部分区域，且每块第一生物特征区域通过对第一生物特征图像分块得到。
在本发明实施例中，第一生物特征图像是用于对用户身份进行识别的基础，其可以人脸图像、虹膜图像和指纹图像中的任意一种。并且人脸图像和虹膜图像可以通过电子设备中的摄像装置获取，而指纹图像则可以通过电子设备中的指纹采集装置获取。在第一生物特征图像是虹膜图像或指纹图像的情况下，可以直接对第一生物特征图像进行分块；在第一生物特征图像是人脸图像的情况下，则需要首先识别出人脸图像中的特征对象所在区域，如人眼、嘴巴和鼻子所在区域，然后再对特征对象所在区域分块以得到多块第一生物特征区域。
在对上述第一生物特征图像进行分块时，可以按照预设分块数量对第一生物特征图像进行分块，即将第一生物特征图像划分成预设分块数量的第一生物特征图像，如图2所示，其中图2左边为一第一生物特征图像，将该第一生物特征图像划分成图2右边所示的n块第一生物特征区域，n为预设分块数量。
在本发明实施例中，预设分块数量可以是一固定取值或者其可以根据识
别准确度来设定，当识别准确度较高时可以增大预设分块数量的取值；当识别准确度较低时则可以降低预设分块数量的取值。此外上述预设分块数量还可以与第一生物特征图像的质量相关，即第一生物特征图像的质量较高时可以增大预设分块数量的取值；当第一生物特征图像的质量较低时则可以降低预设分块数量的取值，其中第一生物特征图像的质量与获取第一生物特征图像的装置的配置参数相关，以摄像装置为例，当摄像装置的分辨率较高时，得到的第一生物特征图像的清晰度越好，则说明第一生物特征图像的质量较高；当摄像装置的分辨率较低时，得到的第一生物特征图像的清晰度则会降低，说明第一生物特征图像的质量较低。
102：对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据，即本发明实施例可以分块来提取第一生物特征图像中的特征数据，与现有技术中直接从整幅第一生物特征图像中的提取不同，尤其是电子设备可以并行处理数据的情况下，采用分块提取方式可以使得电子设备并行对多块第一生物特征区域进行特征提取，从而加快提取速度。
103：对每块第一生物特征区域的特征数据进行特征匹配，其中特征匹配过程是：将第i块第一生物特征区域的特征数据与预先存储的第i块第二生物特征区域的特征数据进行匹配，得到第i块第一生物特征区域的匹配结果，1≤i≤n，n为得到的第一生物特征区域总数，预先存储的第i块第二生物特征区域是对预先存储的合法用户的第二生物特征图像进行分块得到的一个区域。
在本发明实施例中，为能够对每块第一生物特征区域的特征数据进行特征匹配，第二生物特征图像的第二生物特征区域的数量与第一生物特征图像的第一生物特征区域的数量相同，这样才能够对第一生物特征区域和第二生物特征区域进行一对一匹配，其中第一生物特征区域和第二生物特征区域进行一对一匹配的过程如图3所示，即对于第1块至第n块来说，将第1块第一生物特征区域与第1块第二生物特征区域进行匹配、第2块第一生物特征区域与第2块第二生物特征区域进行匹配、......、第n块第一生物特征区域与
第n块第二生物特征区域进行匹配，得到的匹配结果依次为σ1、σ2、......、σn，其中匹配结果以匹配度来表示，其取值范围为0％到100％，且图3中的∈表示特征匹配。
104：基于第一生物特征区域的匹配结果，得到第一生物特征图像的匹配度。
105：基于第一生物特征图像的匹配度，得到第一生物特征图像对应的用户身份识别结果。
其中第一生物特征图像的匹配度用于指示第一生物特征图像相对于第二生物特征图像的匹配情况，如果第一生物特征图像的匹配度指示第一生物特征图像与第二生物特征图像匹配，则表明第一生物特征图像对应的用户身份为合法用户身份；如果第一生物特征图像的匹配度指示第一生物特征图像与第二生物特征图像不匹配，则表明第一生物特征图像对应的用户身份为非法用户身份。
从上述技术方案可知，本发明实施例提供的生物特征身份识别方法在获得第一生物特征图像后，对第一生物特征图像进行分块得到多块第一生物特征区域，再对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据，然后基于第一生物特征区域的匹配结果，得到第一生物特征图像的匹配度，并基于第一生物特征图像的匹配度，得到第一生物特征图像对应的用户身份识别结果，也就是说本发明可以基于第一生物特征区域的合法匹配的比例来得到第一生物特征图像对应的用户身份识别结果，在一定程度上降低对第一生物特征图像的质量要求，进而降低因拍摄角度、光照等导致采集第一生物特征图像不完整导致的重复识别的概率。
在本发明实施例中，第一生物特征图像的匹配度可以通过对各个第一生物特征区域的匹配结果进行归一化处理得到，其归一化处理过程如下：
首先，设置第一生物特征区域的判断阈值，记为θ，其中θ的取值可以根据电子设备的安全等级要求而定，例如在安全等级要求较低应用下θ可以取较小的值，比如60％，相反对安全等级要求较高的应用下θ可以取80％到90％。
用Qi表示每块第一生物特征区域的匹配情况，其取值如下：
其次，计算匹配的第一生物特征区域的数量在分块总数中的比例，记为η，取值如下：
以下结合具体实例对本发明实施例提供的技术方案进行详细说明：假设第一生物特征图像为虹膜图像，并将虹膜图像分为上下两个区域，即n＝2，分别提取上下两个区域的特征。这两个区域的特征数据和第二生物特征图像的特征数据为：
上半区域特征数据＝{50、23、32、41、65、85、49、76、21、11}；
下半区域特征数据＝{47、35、47、38、69、85、25、35、99、52}；
第二生物特征图像的上半区域特征序列＝{50、78、32、41、65、85、49、55、21、11}；
第二生物特征图像的下半区域特征序列＝{33、45、47、77、39、41、25、85、22、58}；
两个区域的匹配流程如图4所示，其中粗体部分的数据表示不匹配，正常字体部分的数据匹配，且上半区域的匹配度为80％，由于虹膜图像受光照和拍摄角度影响导致下半区域采集的图像质量较差，使得下半区域提取的特征不够准确，最终导致下半区域的匹配度为20％。若是按照直接从整幅虹膜
图像进行特征提取并匹配，其匹配度最高为50％。若匹配的阈值θ＝70％，则现有直接从整幅虹膜图像进行特征提取并匹配的识别结果是不匹配，即用户身份为非法用户身份，这就造成合法用户的误判；而采用本发明实施例得到分块识别的结果是上半区域匹配，下半区域不匹配。若选取η＝40％，则最终的识别结果是匹配，正确识别合法用户的身份。
此外本发明实施例提供的生物特征身份识别方法还可以串行对每块第一生物特征区域的特征数据进行匹配，即对于第1块第一生物特征区域至第n块第一生物特征区域来说，依次对每块第一生物特征区域的特征数据进行匹配，这样在对预设数量的第一生物特征区域进行匹配后，判断基于预设数量的第一生物特征区域的匹配结果得到的第一生物特征图像的匹配度是否指示第一生物特征图像对应的用户身份识别结果合法，如果是，则终止对剩余的第一生物特征区域的匹配；如果否可以在对剩余的第一生物特征区域依次进行匹配的过程中，判断每次匹配后得到的第一生物特征图像的匹配度是否指示第一生物特征图像对应的用户身份识别结果合法，如果是，则终止对剩余的第一生物特征区域的匹配，如果否则继续匹配并判断。
也就是说在依次对每块第一生物特征区域的特征数据进行匹配的情况下，可以首先对部分第一生物特征区域进行匹配，然后得到基于这部分第一生物特征区域的第一生物特征图像的匹配度，若此时得到的第一生物特征图像的匹配度指示第一生物特征图像对应的用户身份识别结果合法，则可以不再对剩余的第一生物特征区域进行匹配，从而减少匹配的第一生物特征区域的数量，进而可以节省系统资源。其中预设数量可以大于等于n/2，具体取值本发明实施例并不加以限定。
在这里需要说明的一点是：本发明实施例提供的第一生物特征图像的匹配度与匹配的第一生物特征区域的数量相关，若一百块第一生物特征区域中只有一块匹配，即使其匹配度达到100％，也判定用户身份为非法用户身份，以排除因提取的特征数据不精确导致误识别。若是要求匹配的第一生物特征
区域过多则会失去分块的意义，因此η的取值在设置过程中不宜较高，例如全等级要求较高的应用下η最大取值可以是70％，安全等级要求不高的应用下可选取稍微小的比例比如30％。
与上述方法实施例相对应的，本发明实施例还提供一种生物特征身份识别装置，其结构示意图如图5所示，可以包括：分块单元11、提取单元12、匹配单元13、计算单元14和识别单元15。
分块单元11，用于在获得第一生物特征图像后，对第一生物特征图像进行分块，得到多块第一生物特征区域。也就是说每块第一生物特征区域是第一生物特征图像的部分区域，且每块第一生物特征区域通过对第一生物特征图像分块得到。
在本发明实施例中，第一生物特征图像是用于对用户身份进行识别的基础，其可以人脸图像、虹膜图像和指纹图像中的任意一种。并且人脸图像和虹膜图像可以通过电子设备中的摄像装置获取，而指纹图像则可以通过电子设备中的指纹采集装置获取。在第一生物特征图像是虹膜图像或指纹图像的情况下，可以直接对第一生物特征图像进行分块；在第一生物特征图像是人脸图像的情况下，则需要首先识别出人脸图像中的特征对象所在区域，如人眼、嘴巴和鼻子所在区域，然后再对特征对象所在区域分块以得到多块第一生物特征区域。
在对上述第一生物特征图像进行分块时，分块单元11可以按照预设分块数量对第一生物特征图像进行分块，即将第一生物特征图像划分成预设分块数量的第一生物特征图像，如图2所示，其中图2左边为一第一生物特征图像，将该第一生物特征图像划分成图2右边所示的n块第一生物特征区域，n为预设分块数量。
在本发明实施例中，预设分块数量可以是一固定取值或者其可以根据识别准确度来设定，当识别准确度较高时可以增大预设分块数量的取值；当识别准确度较低时则可以降低预设分块数量的取值。此外上述预设分块数量还
可以与第一生物特征图像的质量相关，即第一生物特征图像的质量较高时可以增大预设分块数量的取值；当第一生物特征图像的质量较低时则可以降低预设分块数量的取值，其中第一生物特征图像的质量与获取第一生物特征图像的装置的配置参数相关，以摄像装置为例，当摄像装置的分辨率较高时，得到的第一生物特征图像的清晰度越好，则说明第一生物特征图像的质量较高；当摄像装置的分辨率较低时，得到的第一生物特征图像的清晰度则会降低，说明第一生物特征图像的质量较低。
提取单元12，用于对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据，即本发明实施例可以分块来提取第一生物特征图像中的特征数据，与现有技术中直接从整幅第一生物特征图像中的提取不同，尤其是电子设备可以并行处理数据的情况下，采用分块提取方式可以使得提取单元12并行对多块第一生物特征区域进行特征提取，从而加快提取速度。
匹配单元13，用于对每块第一生物特征区域的特征数据进行特征匹配，其中特征匹配过程是：将第i块第一生物特征区域的特征数据与预先存储的第i块第二生物特征区域的特征数据进行匹配，得到第i块第一生物特征区域的匹配结果，1≤i≤n，n为得到的第一生物特征区域总数，预先存储的第i块第二生物特征区域是对预先存储的合法用户的第二生物特征图像进行分块得到的一个区域。
在本发明实施例中，为能够对每块第一生物特征区域的特征数据进行特征匹配，第二生物特征图像的第二生物特征区域的数量与第一生物特征图像的第一生物特征区域的数量相同，这样才能够对第一生物特征区域和第二生物特征区域进行一对一匹配，其中第一生物特征区域和第二生物特征区域进行一对一匹配的过程如图3所示，即对于第1块至第n块来说，将第1块第一生物特征区域与第1块第二生物特征区域进行匹配、第2块第一生物特征区域与第2块第二生物特征区域进行匹配、......、第n块第一生物特征区域与第n块第二生物特征区域进行匹配，得到的匹配结果依次为σ1、σ2、......、σn，
其中匹配结果以匹配度来表示，其取值范围为0％到100％，且图3中的∈表示特征匹配。
计算单元14，用于基于第一生物特征区域的匹配结果，得到第一生物特征图像的匹配度。
识别单元15，用于基于第一生物特征图像的匹配度，得到第一生物特征图像对应的用户身份识别结果。
其中第一生物特征图像的匹配度用于指示第一生物特征图像相对于第二生物特征图像的匹配情况，如果第一生物特征图像的匹配度指示第一生物特征图像与第二生物特征图像匹配，则表明第一生物特征图像对应的用户身份为合法用户身份；如果第一生物特征图像的匹配度指示第一生物特征图像与第二生物特征图像不匹配，则表明第一生物特征图像对应的用户身份为非法用户身份。
从上述技术方案可知，本发明实施例提供的生物特征身份识别装置在获得第一生物特征图像后，对第一生物特征图像进行分块得到多块第一生物特征区域，再对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据，然后基于第一生物特征区域的匹配结果，得到第一生物特征图像的匹配度，并基于第一生物特征图像的匹配度，得到第一生物特征图像对应的用户身份识别结果。也就是说本发明可以基于第一生物特征区域的合法匹配的比例来得到第一生物特征图像对应的用户身份识别结果，在一定程度上降低对第一生物特征图像的质量要求，进而降低因拍摄角度、光照等导致采集第一生物特征图像不完整导致的重复识别的概率。
在本发明实施例中，第一生物特征图像的匹配度可以通过对各个第一生物特征区域的匹配结果进行归一化处理得到，其归一化处理过程如下：
首先，设置第一生物特征区域的判断阈值，记为θ，其中θ的取值可以根据电子设备的安全等级要求而定，例如在安全等级要求较低应用下θ可以取较小的值，比如60％，相反对安全等级要求较高的应用下θ可以取80％到90％。用Qi表示每块第一生物特征区域的匹配情况，其取值如下：
其次，计算匹配的第一生物特征区域的数量在分块总数中的比例，记为η，
其中n为分块总数，η表示为第一生物特征图像的匹配度。也就是说本发明实施例中第一生物特征图像的匹配度与匹配的第一生物特征区域的数量相关，在因拍摄角度、光照等导致第一生物特征图像不完整的情况下，本发明实施例可以通过划分更多的第一生物特征区域的方式来提高与第二生物特征区域匹配的第一生物特征区域的数量，以此提高相对于基于整幅第一生物特征图像的一次匹配方式的识别准确度。
此外本发明实施例提供的生物特征身份识别装置还具有其他结构，如图6所示，其中图6所示的生物特征身份识别装置中提取单元12串行对每块第一生物特征区域的特征数据进行匹配，即对于第1块第一生物特征区域至第n块第一生物特征区域来说，提取单元12依次对每块第一生物特征区域的特征数据进行匹配。
判断单元16，用于在对预设数量的第一生物特征区域进行匹配后，判断基于预设数量的第一生物特征区域的匹配结果得到的第一生物特征图像的匹配度是否指示第一生物特征图像对应的用户身份识别结果合法，如果是，则触发提取单元12终止对剩余的第一生物特征区域的匹配；如果否则触发提取单元12在对剩余的第一生物特征区域依次进行匹配，并在提取单元对剩余的第一生物特征区域依次进行匹配的过程中，判断单元16仍需要判断每次匹配后得到的第一生物特征图像的匹配度是否指示第一生物特征图像对应的用户身份识别结果合法，如果是，则触发提取单元12终止对剩余的第一生物特征区域的匹配，如果否则触发提取单元12继续匹配并判断。
也就是说在依次对每块第一生物特征区域的特征数据进行匹配的情况
下，可以首先对部分第一生物特征区域进行匹配，然后得到基于这部分第一生物特征区域的第一生物特征图像的匹配度，若此时得到的第一生物特征图像的匹配度指示第一生物特征图像对应的用户身份识别结果合法，则可以不再对剩余的第一生物特征区域进行匹配，从而减少匹配的第一生物特征区域的数量，进而可以节省系统资源。其中预设数量可以大于等于n/2，具体取值本发明实施例并不加以限定。
在这里需要说明的一点是：本发明实施例提供的第一生物特征图像的匹配度与匹配的第一生物特征区域的数量相关，若一百块第一生物特征区域中只有一块匹配，即使其匹配度达到100％，也判定用户身份为非法用户身份，以排除因提取的特征数据不精确导致误识别。若是要求匹配的第一生物特征区域过多则会失去分块的意义，因此η的取值在设置过程中不宜较高，例如全等级要求较高的应用下η最大取值可以是70％，安全等级要求不高的应用下可选取稍微小的比例比如30％。
需要说明的是，在本文中，诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
对所公开的实施例的上述说明，使本领域技术人员能够实现或使用本发明。对这些实施例的多种修改对本领域技术人员来说将是显而易见的，本文中所定义的一般原理可以在不脱离本发明的精神或范围的情况下，在其它实施例中实现。因此，本发明将不会被限制于本文所示的这些实施例，而是要符合与本文所公开的原理和新颖特点相一致的最宽的范围。
以上所述仅是本发明的优选实施方式，应当指出，对于本技术领域的普
通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。
Claims (10)
- 一种生物特征身份识别方法，其特征在于，所述方法包括：在获得第一生物特征图像后，对所述第一生物特征图像进行分块，得到多块第一生物特征区域；对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据；对每块第一生物特征区域的特征数据进行特征匹配，其中特征匹配过程是：将第i块第一生物特征区域的特征数据与预先存储的第i块第二生物特征区域的特征数据进行匹配，得到第i块第一生物特征区域的匹配结果，1≤i≤n，n为得到的第一生物特征区域总数，所述预先存储的第i块第二生物特征区域是对预先存储的合法用户的第二生物特征图像进行分块得到的一个区域；基于所述第一生物特征区域的匹配结果，得到所述第一生物特征图像的匹配度，其中所述第一生物特征图像的匹配度用于指示所述第一生物特征图像相对于所述第二生物特征图像的匹配情况；基于所述第一生物特征图像的匹配度，得到所述第一生物特征图像对应的用户身份识别结果。
- 根据权利要求1所述的方法，其特征在于，所述在获得第一生物特征图像后，对所述第一生物特征图像进行分块，得到多块第一生物特征区域，包括：在获得第一生物特征图像后，按照预设分块数量对所述第一生物特征图像进行分块，得到多块第一生物特征区域。
- 根据权利要求1所述的方法，其特征在于，所述对每块第一生物特征区域的特征数据进行特征匹配，包括：并行对多块第一生物特征区域的特征数据进行匹配。
- 根据权利要求1所述的方法，其特征在于，所述对每块第一生物特征区域的特征数据进行特征匹配，包括：依次对每块第一生物特征区域的特征数据进行匹配。
- 根据权利要求4所述的方法，其特征在于，所述方法还包括：在对预 设数量的第一生物特征区域进行匹配后，判断基于预设数量的第一生物特征区域的匹配结果得到的第一生物特征图像的匹配度是否指示所述第一生物特征图像对应的用户身份识别结果合法；如果是，则终止对剩余的第一生物特征区域的匹配。
- 一种生物特征身份识别装置，其特征在于，所述装置包括：分块单元，用于在获得第一生物特征图像后，对所述第一生物特征图像进行分块，得到多块第一生物特征区域；提取单元，用于对每块第一生物特征区域进行特征提取，得到每块第一生物特征区域的特征数据；匹配单元，用于对每块第一生物特征区域的特征数据进行特征匹配，其中特征匹配过程是：将第i块第一生物特征区域的特征数据与预先存储的第i块第二生物特征区域的特征数据进行匹配，得到第i块第一生物特征区域的匹配结果，1≤i≤n，n为得到的第一生物特征区域总数，所述预先存储的第i块第二生物特征区域是对预先存储的合法用户的第二生物特征图像进行分块得到的一个区域；计算单元，用于基于所述第一生物特征区域的匹配结果，得到所述第一生物特征图像的匹配度，其中所述第一生物特征图像的匹配度用于指示所述第一生物特征图像相对于所述第二生物特征图像的匹配情况；识别单元，用于基于所述第一生物特征图像的匹配度，得到所述第一生物特征图像对应的用户身份识别结果。
- 根据权利要求6所述的装置，其特征在于，所述分块单元用于：在获得第一生物特征图像后，按照预设分块数量对所述第一生物特征图像进行分块，得到多块第一生物特征区域。
- 根据权利要求6所述的装置，其特征在于，所述提取单元用于：并行对多块第一生物特征区域的特征数据进行匹配。
- 根据权利要求6所述的装置，其特征在于，所述提取单元用于：依次对每块第一生物特征区域的特征数据进行匹配。
- 根据权利要求9所述的装置，其特征在于，所述装置还包括：判断单元，用于在对预设数量的第一生物特征区域进行匹配后，判断基于预设数量的第一生物特征区域的匹配结果得到的第一生物特征图像的匹配度是否指示所述第一生物特征图像对应的用户身份识别结果合法，如果是，则触发所述提取单元终止对剩余的第一生物特征区域的匹配。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510626303.9A CN105631285A (zh) | 2015-09-28 | 2015-09-28 | 一种生物特征身份识别方法及装置 |
| CN201510626303.9 | 2015-09-28 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2017054276A1 true WO2017054276A1 (zh) | 2017-04-06 |
Family
ID=56046211
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2015/093379 WO2017054276A1 (zh) | 2015-09-28 | 2015-10-30 | 一种生物特征身份识别方法及装置 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN105631285A (zh) |
| WO (1) | WO2017054276A1 (zh) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116346489A (zh) * | 2023-04-14 | 2023-06-27 | 江苏奥诺电能科技有限公司 | 一种互联网信息安全通信系统 |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106886742A (zh) * | 2015-12-16 | 2017-06-23 | 东莞市中控电子技术有限公司 | 一种虹膜采集方法及虹膜采集装置 |
| CN107229915A (zh) * | 2017-05-26 | 2017-10-03 | 北京小米移动软件有限公司 | 生物特征识别方法、装置、设备及存储介质 |
| CN107358699B (zh) * | 2017-07-17 | 2020-04-24 | 深圳市斑点猫信息技术有限公司 | 一种安全验证方法及系统 |
| CN108196740B (zh) * | 2017-11-28 | 2019-08-20 | 维沃移动通信有限公司 | 一种图标显示方法、装置及移动终端 |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101122949A (zh) * | 2007-08-30 | 2008-02-13 | 中国科学技术大学 | 一种基于局部频率特征和局部方向特征的虹膜识别方法 |
| CN103150561A (zh) * | 2013-03-19 | 2013-06-12 | 华为技术有限公司 | 人脸识别方法和设备 |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101231695A (zh) * | 2008-01-19 | 2008-07-30 | 电子科技大学中山学院 | 一种基于多分辨率分析的虹膜识别方法 |
| CN101916363B (zh) * | 2010-05-28 | 2012-06-20 | 深圳大学 | 一种虹膜特征设计与编码方法及虹膜识别系统 |
| CN101923640B (zh) * | 2010-08-04 | 2013-03-20 | 北京中科虹霸科技有限公司 | 基于鲁棒纹理特征和机器学习对伪造虹膜图像判别的方法 |
| CN102542243A (zh) * | 2010-12-17 | 2012-07-04 | 北京理工大学 | 一种基于lbp图像和分块编码的虹膜特征提取方法 |
| CN102087705B (zh) * | 2011-03-16 | 2012-07-25 | 山东大学 | 一种基于毯子维和缺项的虹膜识别方法 |
| CN102521575B (zh) * | 2011-12-16 | 2014-07-02 | 北京天诚盛业科技有限公司 | 基于多方向Gabor和Adaboost虹膜识别方法 |
| CN102867308B (zh) * | 2012-09-11 | 2015-06-03 | 大连理工大学 | 一种电脑输出视频图像变化检测的方法 |
-
2015
- 2015-09-28 CN CN201510626303.9A patent/CN105631285A/zh active Pending
- 2015-10-30 WO PCT/CN2015/093379 patent/WO2017054276A1/zh active Application Filing
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101122949A (zh) * | 2007-08-30 | 2008-02-13 | 中国科学技术大学 | 一种基于局部频率特征和局部方向特征的虹膜识别方法 |
| CN103150561A (zh) * | 2013-03-19 | 2013-06-12 | 华为技术有限公司 | 人脸识别方法和设备 |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116346489A (zh) * | 2023-04-14 | 2023-06-27 | 江苏奥诺电能科技有限公司 | 一种互联网信息安全通信系统 |
| CN116346489B (zh) * | 2023-04-14 | 2023-10-10 | 江苏奥诺电能科技有限公司 | 一种互联网信息安全通信系统 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105631285A (zh) | 2016-06-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10043061B2 (en) | Facial matching system | |
| WO2019232862A1 (zh) | 嘴巴模型训练方法、嘴巴识别方法、装置、设备及介质 | |
| WO2019232866A1 (zh) | 人眼模型训练方法、人眼识别方法、装置、设备及介质 | |
| WO2017054276A1 (zh) | 一种生物特征身份识别方法及装置 | |
| KR100374708B1 (ko) | 회전영상의 보정에 의한 비접촉식 홍채인식방법 | |
| WO2021036436A1 (zh) | 一种人脸识别方法及装置 | |
| WO2019153739A1 (zh) | 基于人脸识别的身份认证方法、装置、设备和存储介质 | |
| WO2018216629A1 (ja) | 情報処理装置、情報処理方法、及びプログラム | |
| CN107844736B (zh) | 虹膜定位方法和装置 | |
| JP2018198053A (ja) | 情報処理装置、情報処理方法、及びプログラム | |
| WO2020253062A1 (zh) | 一种图像边界的检测方法及装置 | |
| WO2020244071A1 (zh) | 基于神经网络的手势识别方法、装置、存储介质及设备 | |
| WO2019061658A1 (zh) | 眼镜定位方法、装置及存储介质 | |
| US9489561B2 (en) | Method and system for estimating fingerprint pose | |
| KR101632912B1 (ko) | 지문 인식을 이용한 사용자 인증 방법 | |
| US8090151B2 (en) | Face feature point detection apparatus and method of the same | |
| Abate et al. | BIRD: Watershed based iris detection for mobile devices | |
| US20130202159A1 (en) | Apparatus for real-time face recognition | |
| Thalji et al. | Iris Recognition using robust algorithm for eyelid, eyelash and shadow avoiding | |
| CN112434578A (zh) | 口罩佩戴规范性检测方法、装置、计算机设备和存储介质 | |
| WO2021000832A1 (zh) | 匹配人脸的方法、装置、计算机设备和存储介质 | |
| Hughes et al. | Detection of contact-lens-based iris biometric spoofs using stereo imaging | |
| Pundlik et al. | Iris segmentation in non-ideal images using graph cuts | |
| WO2019042195A1 (zh) | 一种人体目标身份识别方法及装置 | |
| CN105279492A (zh) | 虹膜识别的方法和装置 |
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: 15905188 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: 15905188 Country of ref document: EP Kind code of ref document: A1 |