CN111325156B - Face recognition method, device, equipment and storage medium - Google Patents
Face recognition method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN111325156B CN111325156B CN202010112341.3A CN202010112341A CN111325156B CN 111325156 B CN111325156 B CN 111325156B CN 202010112341 A CN202010112341 A CN 202010112341A CN 111325156 B CN111325156 B CN 111325156B
- Authority
- CN
- China
- Prior art keywords
- face
- image
- category
- sub
- determining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
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
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/55—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The embodiment of the invention discloses a face recognition method, a face recognition device, face recognition equipment and a storage medium. The method comprises the following steps: acquiring face characteristic information of a face image to be identified, and determining a target face category to which the face image to be identified belongs by using a preset face classification model based on the face characteristic information; determining face images to be compared from a face classification database according to the target face types; and comparing the face image to be identified with the face image to be compared to obtain a face identification result. Through the technical scheme, the face recognition efficiency is improved.
Description
Technical Field
The embodiment of the invention relates to computer technology, in particular to a face recognition method, a device, equipment and a storage medium.
Background
With the development of technology, the off-line payment mode has occurred in face recognition payment, that is, the face image of the payer is scanned by the collection device during payment, and the face information and the personal payment account corresponding to the payer in the local database are determined by recognizing the face image, so that the off-line payment is completed. In the face recognition payment process, the efficiency and accuracy of face recognition largely determine the efficiency and accuracy of face recognition payment.
In the existing face recognition technology, after the number of face images in a database exceeds a certain critical value, the face recognition efficiency is reduced due to the rapid increase of the comparison number of the face features. Therefore, in the face recognition payment system, the number of face images stored in the local database is generally less than or equal to a critical value to ensure payment efficiency. However, in an actual payment scene, the number of actual face images is far more than that in a local database due to the large mobility of personnel.
At present, in order to solve the problems of a large number of face images and payment efficiency in a payment scene, the following manner is generally adopted: the number of face images in a local database is increased, face information and a personal payment account are locked through mobile phone numbers, the face images are prevented from being compared one by one, and the payment efficiency is improved; or, keeping the number of the face images in the local database unchanged, and dynamically changing the face images in the database (namely dynamically constructing the regional face database) so that the face images in the local database can adapt to the payment scene requirement.
In the process of implementing the present invention, the inventor finds that at least the following problems exist in the prior art: (1) The face recognition payment mode through the mobile phone number increases the number of face images in a local database, so that the scene applicability of the face recognition payment is improved, but the convenience of the face recognition payment is reduced due to the introduction of the mobile phone number; (2) The method for dynamically constructing the regional face database does not essentially increase the number of face images in the local face database, and the dynamic construction process of the regional face database has certain time loss and extra performance loss, so that the payment efficiency of face recognition is reduced to a certain extent.
Disclosure of Invention
The embodiment of the invention provides a face recognition method, a face recognition device, face recognition equipment and a storage medium, so that face recognition efficiency is improved, and scene applicability and convenience of face recognition payment are improved.
In a first aspect, an embodiment of the present invention provides a face recognition method, including:
acquiring face characteristic information of a face image to be identified, and determining a target face category to which the face image to be identified belongs by using a preset face classification model based on the face characteristic information;
determining face images to be compared from a face classification database according to the target face types;
and comparing the face image to be identified with the face image to be compared to obtain a face identification result.
In a second aspect, an embodiment of the present invention further provides a face recognition device, including:
the target face type determining module is used for acquiring face characteristic information of a face image to be recognized and determining a target face type of the face image to be recognized by utilizing a preset face classification model based on the face characteristic information;
the face image to be compared determining module is used for determining face images to be compared from a face classification database according to the target face category;
And the face recognition result obtaining module is used for obtaining a face recognition result by comparing the face image to be recognized with the face image to be compared.
In a third aspect, an embodiment of the present invention further provides an electronic device, including:
one or more processors;
storage means for storing one or more programs,
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the face recognition method provided by any embodiment of the present invention.
In a fourth aspect, an embodiment of the present invention further provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor implements the face recognition method provided by any embodiment of the present invention.
According to the embodiment of the invention, the face characteristic information of the face image to be recognized is obtained, and the target face category to which the face image to be recognized belongs is determined by utilizing a preset face classification model based on the face characteristic information; determining face images to be compared from a face classification database according to the target face types; and comparing the face image to be recognized with the face image to be compared to obtain a face recognition result. The method and the device realize that the local face database is divided into a plurality of different face categories, face recognition is carried out in a certain face category, and the problem that the face recognition efficiency is reduced due to the fact that the total number of face images in the local face database exceeds the number of face images when the face recognition efficiency begins to be reduced is avoided, so that the total number of face images in the local face database can be increased, the face recognition efficiency is improved, and the applicability of face recognition payment to an actual payment scene is improved on the basis of face recognition payment convenience.
Drawings
Fig. 1 is a flowchart of a face recognition method according to a first embodiment of the present invention;
fig. 2 is a flowchart of a face recognition method in the second embodiment of the present invention;
FIG. 3 is a schematic diagram of a partial binary pattern algorithm according to a second embodiment of the present invention;
FIG. 4 is a search index representing intent of image search in accordance with a second embodiment of the present invention;
fig. 5 is a schematic structural diagram of a face recognition device according to a third embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention.
Detailed Description
The invention is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting thereof. It should be further noted that, for convenience of description, only some, but not all of the structures related to the present invention are shown in the drawings.
Example 1
The face recognition method provided by the embodiment can be suitable for the situation that the face recognition technology is needed, such as face recognition payment. The method may be performed by a face recognition device, which may be implemented in software and/or hardware, and which may be integrated in an electronic device having an image processing function, such as a notebook computer, a desktop computer, a server, a cash register terminal device, or the like. Referring to fig. 1, the method of this embodiment specifically includes the following steps:
S110, acquiring face characteristic information of a face image to be identified, and determining a target face category to which the face image to be identified belongs by using a preset face classification model based on the face characteristic information.
The face image to be identified refers to an image to be identified by using a face identification technology, and the image can be obtained by scanning a face in real time or can be read from a local or cloud storage space. When the application scene of the face recognition is face recognition payment, the face image to be recognized is obtained by real-time scanning. The face feature information is information describing a face feature in a face image, and may be, for example, digital information such as a face feature matrix or a face feature vector. The expression form of the face characteristic information is consistent with the model input form of a preset face classification model. The preset face classification model refers to a model which is trained in advance and used for classifying face images. Illustratively, the preset face classification model is obtained by training at least one supervised classification machine learning model in advance. Training samples composed of face categories and corresponding face feature information can be collected, and model training is carried out on 1 or more supervised and classified machine learning models by using the training samples, so that a preset face classification model is obtained. The supervised classification machine learning model may be, for example, a decision tree, bayes, artificial neural network, or k-nearest neighbor classification model, etc. The advantage of this arrangement is that the classification accuracy of the face image can be improved. The face category is a category determined by classifying the face of the user, and may be, for example, a male face, a male round face, a male melon seed face, a female round face, a female melon seed face, and the like, which may be defined according to the service.
Specifically, after the face image to be recognized is obtained, feature extraction is performed on the face image to be recognized by using a feature extraction algorithm such as principal component analysis (principal components analysis, PCA), and the result of feature extraction is face feature information of the face image to be recognized. And then, inputting the face characteristic information into a preset face classification model, and obtaining a model output result through model operation, wherein the model output result is the face type (namely the target face type) corresponding to the face image to be recognized.
S120, determining the face images to be compared from a face classification database according to the target face types.
The face classification database is a database for classifying and storing different face images according to face categories. The face image to be compared is a face image for matching with the face image to be recognized, which may be one or more face images.
Specifically, in order to ensure the face recognition efficiency in the related art, the total number of face images in the face database is limited not to exceed the corresponding number of face images (i.e., the number of face images in the preset database) when the face recognition efficiency begins to decrease (i.e., the decreasing inflection point of the face recognition rate), that is, the limited number of face images in the face database may result in some face recognition without results or results errors, and for face recognition payment applications, it cannot accommodate some face images appearing in the actual face recognition payment scene, thereby resulting in face recognition payment failure. In the embodiment of the invention, in order to improve the face recognition efficiency and accuracy, the face image capacity in the face database is expanded, the face database is converted into the face classification database, the face image capacity under each face type can be at most the number of the face images of the preset database, and the subsequent face recognition is carried out under a single face type. In the implementation, according to the target face category corresponding to the face image to be recognized, matching the corresponding face category from a face classification database, and determining the face image to be compared by utilizing each face image contained in the matched face category.
Illustratively, the face classification database is pre-built by: inputting each face image in a face database into a preset face classification model, and determining the face category to which each face image belongs; and determining face images contained in the corresponding face categories from the face images belonging to each face category according to the number of the face images belonging to each face category and the number of the face images of a preset database, and constructing a face classification database, wherein the number of the face images of the preset database is the number of the face images at the falling inflection point of the face recognition rate. The process for constructing the face classification database is as follows: each face image in the face database is input into a preset face classification model, and the face classification of the corresponding face image is determined by the model output result. At this time, a certain number of face images are corresponding to each face category. And then comparing the number of the face images contained in each face category with the number of the face images of a preset database, and determining the face images contained in each face category in the face classification database according to the comparison result. For example, if the number of face images belonging to a certain face class in the face database is less than or equal to the number of face images in the preset database, determining all face images belonging to the face class as face images contained in the face class in the face classification database; if the number of the face images belonging to a certain face class in the face database is larger than the number of the face images of the preset database, the face images with the number of the face images of the preset database are screened out from all the face images belonging to the face class, and the face images are determined to be the face images contained in the face class in the face classification database. According to the process, a face classification database can be constructed from the face database. The face recognition method has the advantages that the face image capacity in the face database can be expanded to a certain extent, and the face recognition efficiency can be improved.
Illustratively, determining face images to be compared from a face classification database according to target face categories includes:
A. matching the target face category with each face category in the face classification database, determining the matched face category, and determining the sub-category face image from the face classification database according to the matched face category.
The sub-category face images refer to face images in a certain face category.
Specifically, matching the target face category with each face category in the face classification database to obtain a face category consistent in matching as a matching face category. And then taking the face images contained in the matching face categories in the face classification database as sub-category face images.
B. And carrying out image retrieval on the facial images of the sub-categories according to the facial images to be identified, and determining the facial images to be compared from the facial images of the sub-categories according to the image retrieval result.
Specifically, the number of facial images of sub-categories included in the matching facial category is relatively large, and in this embodiment, in order to further reduce the number of facial comparison times in the face recognition process, an image retrieval technology is used to perform image screening on all the facial images of sub-categories determined in the step a. In specific implementation, a text-based image retrieval technology or a content-based image retrieval technology can be adopted, the face images to be identified are taken as retrieval references, and all sub-category face images contained in the matched face category are retrieved to obtain an image retrieval result, such as a plurality of sub-category face images meeting the retrieval conditions in an image retrieval algorithm. And then, screening all the facial images of the sub-categories contained in the matched facial category according to the facial images of the sub-categories in the image retrieval result, so that the facial images to be compared can be determined.
The method has the advantages that the sub-category face images contained in the target face category are screened through the image retrieval technology, the number of face images to be compared is further reduced, face recognition efficiency is further improved, and the payment efficiency of face recognition payment is further improved.
S130, comparing the face image to be recognized with the face image to be compared to obtain a face recognition result.
Specifically, a face recognition algorithm, such as a face recognition algorithm based on geometric features, a face recognition algorithm based on feature faces (PCA), a face recognition algorithm based on a neural network, a face recognition algorithm based on elastic diagram matching or a face recognition algorithm based on a Support Vector Machine (SVM), is adopted, and face images to be compared are utilized to carry out face recognition on the face images to be identified, so that a face recognition result is obtained.
According to the technical scheme, the target face category of the face image to be recognized is determined by acquiring the face characteristic information of the face image to be recognized and utilizing a preset face classification model based on the face characteristic information; determining face images to be compared from a face classification database according to the target face types; and comparing the face image to be recognized with the face image to be compared to obtain a face recognition result. The method and the device realize that the local face database is divided into a plurality of different face categories, face recognition is carried out in a certain face category, and the problem that the face recognition efficiency is reduced due to the fact that the total number of face images in the local face database exceeds the number of face images when the face recognition efficiency begins to be reduced is avoided, so that the total number of face images in the local face database can be increased, the face recognition efficiency is improved, and the applicability of face recognition payment to an actual payment scene is improved on the basis of face recognition payment convenience.
Example two
The present embodiment further optimizes "image retrieval of sub-category face images according to face images to be recognized" based on the above-described first embodiment. On the basis, the face images to be compared can be further optimized from the sub-category face images according to the image retrieval result. Wherein the explanation of the same or corresponding terms as those of the above embodiments is not repeated herein. Referring to fig. 2, the face recognition method provided in this embodiment includes:
s210, acquiring face characteristic information of a face image to be recognized, and determining a target face category to which the face image to be recognized belongs by using a preset face classification model based on the face characteristic information.
S220, matching the target face category with each face category in the face classification database, determining a matched face category, and determining a sub-category face image from the face classification database according to the matched face category.
And S230, extracting features of the face image to be identified, and obtaining a dimension reduction feature matrix of the face image to be identified based on the feature extraction result.
Specifically, binary feature extraction is performed on the face image to be identified, for example, feature extraction is performed on the face image to be identified by using a local binary pattern (Local Binary Pattern, LBP) algorithm, and a feature extraction result, such as an LBP feature matrix, is obtained. The principle of the LBP algorithm is shown in fig. 3, a certain pixel point in the face image is taken as a circle center, and a radius r (see fig. 3 (a)) is taken; taking the circle center pixel value 2 as a threshold value, setting the pixel value of the pixel point which is larger than the threshold value 2 in the circle as 1, otherwise setting the pixel value as 0, and obtaining a result as shown in fig. 3 (b); then, starting from a certain pixel point on the circumference, the pixel points on the circumference are connected in series in a counterclockwise manner, so that 10101011 can be obtained. And processing each region in the face image to be identified according to the process, so that an LBP characteristic matrix of the whole face image to be identified can be obtained.
After the feature extraction result of the image to be identified is obtained, in order to further improve the subsequent image retrieval efficiency, in this embodiment, the feature extraction result needs to be subjected to dimension reduction processing. For example, the LBP feature matrix may be mapped from euclidean space to hamming space, so that the dimension of the matrix may be reduced, and the result obtained by space conversion is the dimension-reduced feature matrix.
S240, determining the vector similarity between each dimension reduction feature vector and at least one preset retrieval feature vector in the dimension reduction feature matrix.
The preset search feature vector refers to a preset vector, which is used for converting the feature vector corresponding to the image into the corresponding text identification information (i.e. the preset search label). Since the feature vectors in the sub-class face image and the face image to be identified use the same preset retrieval feature vector, the number and the value of the elements in the preset retrieval feature vector can be empirically set. Vector similarity refers to the similarity between two vectors, which can be characterized by a Manhattan distance or a Hamming distance between the two vectors, etc.
Specifically, the idea of image retrieval in this embodiment is: and characterizing the feature vector of each dimension (namely the dimension-reduced feature vector) in the dimension-reduced feature matrix of the face image to be identified by using a preset retrieval tag corresponding to the preset retrieval feature vector, and subsequently retrieving by using the preset retrieval tag corresponding to each dimension. In the implementation, the vector similarity between each dimension-reduction feature vector and each preset retrieval feature vector in the dimension-reduction feature matrix is calculated. For example, if a certain dimension-reduction feature vector in the dimension-reduction feature matrix contains 8 elements and a certain preset search feature vector contains 4 elements, 2 vector similarities can be obtained between the dimension-reduction feature vector and the preset search feature vector through calculation. In this way, all vector similarities can be calculated.
S250, determining a search tag sequence of each dimension-reduced feature vector according to the similarity of each vector, a preset similarity threshold and preset search tags corresponding to each preset search feature vector.
The preset similarity threshold is a preset similarity, which is used for judging whether the two vectors are similar vectors or not. In this step, taking the manhattan distance as an example, the smaller the manhattan distance is, the more similar the two vectors are, so the predetermined similarity threshold may be a smaller value of the manhattan distance. The search tag sequence refers to a tag result composed of a plurality of preset search tags.
Specifically, for a certain dimension-reduction feature vector and a certain preset search feature vector, comparing the similarity of each corresponding vector with a preset similarity threshold. If the similarity of a certain vector is smaller than or equal to a preset similarity threshold, the sub-feature vector in the dimension reduction feature vector corresponding to the similarity of the vector and the preset retrieval feature vector are similar vectors, and the sub-feature vector can be characterized by a preset retrieval tag corresponding to the preset retrieval feature vector. The sub feature vector is composed of partial elements in the dimension reduction feature vector, the partial elements are continuous, and the number of the elements is consistent with that of the preset retrieval feature vector. If the vector similarity is greater than a predetermined similarity threshold, the sub-feature vector may not be characterized by a predetermined search tag corresponding to the predetermined search feature vector. According to the process, the search tag sequences corresponding to the dimension reduction feature vectors can be determined, and then the search tag sequences corresponding to all the dimension reduction feature vectors are determined. Thus, the dimension-reduction feature matrix corresponds to a plurality of search tag sequences, one search tag sequence representing each matrix dimension.
S260, determining an image retrieval result according to the retrieval tag sequences of the various dimensions of the face images to be identified and the retrieval tag sequences of the various dimensions of the face images of each sub-category.
Specifically, before this step, each sub-category face image determined in S220 is characterized as a plurality of search tag sequences according to the flow of S230 to S250. Therefore, image retrieval can be carried out on the face images to be identified in each sub-category of face images according to the plurality of retrieval tag sequences corresponding to the dimension reduction feature matrix and the plurality of retrieval tag sequences corresponding to the face images of each sub-category, and an image retrieval result is obtained.
Illustratively, determining the image retrieval result according to the respective dimension retrieval tag sequences of the face images to be identified and the respective dimension retrieval tag sequences of the face images of each sub-category includes: according to the query rule with the same preset retrieval tag in the same vector dimension, taking each dimension retrieval tag sequence of the face image to be recognized as a query standard, carrying out image query on each sub-category face image according to each dimension retrieval tag sequence of each sub-category face image, and determining each queried sub-category face image as an image retrieval result.
Specifically, the vector dimension of the feature vector corresponding to the face image is taken as a retrieval unit, the preset retrieval label is taken as a retrieval index, and an image retrieval index table is established by utilizing a plurality of retrieval label sequences corresponding to each sub-category of face image. Referring to fig. 4, feature vectors corresponding to face images of each sub-category have alpha p Dimension, for each dimension, counting sub-category face images including preset search labels 00, 01, 10 and 11 in the search label sequence respectively, and detecting by presetThe index relationship is established between the index label and the image identification of the sub-category face image, and the index table shown in fig. 4 can be obtained. And then, inquiring the image identification of the sub-category face images corresponding to each preset retrieval tag in each dimension in the dimension-reduction feature matrix of the face images to be identified from the retrieval index table. For example, if the search tag sequence in the first dimension of the dimension-reduction feature matrix includes a preset search tag 00, then the sub-category face images corresponding to all the image identifications corresponding to the preset search tag 00 in the first dimension index table in fig. 4 are determined as the partial image search results of the face images to be identified. Similarly, partial image retrieval results corresponding to each dimension can be obtained, and the partial image retrieval results form image retrieval results of the face image to be identified. This arrangement has the advantage of further improving the image retrieval efficiency.
S270, determining the image similarity between the face image to be identified and each sub-category face image in the image retrieval result.
Specifically, in order to further reduce the number of images to be compared, in this embodiment, sub-category face images in the image search result are further screened, where the screening basis is the image similarity between the sub-category face images and the face images to be identified. Therefore, the image similarity between the face image to be identified and each sub-category face image in the image retrieval result needs to be calculated.
Illustratively, determining the image similarity between the face image to be identified and each sub-category face image in the image retrieval result includes: and determining the image similarity between the face image to be identified and the face image of the corresponding sub-category according to the query times of the face image of each sub-category in the image retrieval result, wherein the query times are the times of the face image of the sub-category in the image retrieval process. Specifically, according to the description of S260, the partial image search results of each preset search tag in each dimension correspond to a batch of sub-category face images, and the sub-category face images in the multiple partial image search results are repeated, so that the number of times of repeating a sub-category face image in the whole image search result is the number of times of querying the sub-category face image. Because the query times of the face images of each sub-category are all obtained in the same image retrieval, the retrieval object (the face image to be identified) and the retrieval base (the retrieval index table constructed by the face images of each sub-category) corresponding to each query time are the same, and the query times of the face images of the sub-category can reflect the image similarity between the face images of the sub-category and the face image to be identified, for example, the greater the query times of the face images of a certain sub-category is, the more similar the face images of the sub-category are to the face image to be identified is. Based on this, in this embodiment, the image similarity between the face image to be recognized and the face images of the sub-category may be defined based on the number of queries. For example, the image similarity may be defined as a ratio of the number of queries of the face images of the sub-category to the number of queries of the face images of all sub-categories in the image retrieval result, so that the image similarity between the face image to be identified and each face image of the sub-category in the image retrieval result may be determined based on the number of queries. This has the advantage that the retrieval information in the image retrieval process can be used to more quickly determine the image similarity.
S280, screening face images to be compared from face images of each sub-category in the image retrieval result according to the similarity of each image.
Specifically, the image retrieval results are further screened according to the image similarity. For example, an image similarity threshold may be set, and sub-category face images corresponding to all image similarities greater than the image similarity threshold are determined as face images to be compared; the image similarity can be arranged in a reverse order, and the sub-category face images corresponding to the preset number of image similarity with the front sequence are selected to be determined as the face images to be compared.
S290, comparing the face image to be recognized with the face image to be compared to obtain a face recognition result.
According to the technical scheme, feature extraction is carried out on the face image to be identified, and a dimension reduction feature matrix of the face image to be identified is obtained based on a feature extraction result; determining the vector similarity between each dimension-reduction feature vector and at least one preset retrieval feature vector in the dimension-reduction feature matrix; determining a retrieval tag sequence of each dimension-reduction feature vector according to the similarity of each vector, a preset similarity threshold value and a preset retrieval tag corresponding to each preset retrieval feature vector; and determining an image retrieval result according to the retrieval tag sequences of the dimensions of the face images to be identified and the retrieval tag sequences of the dimensions of the face images of each sub-category. The quick retrieval of the images is realized, and the determination efficiency and the accuracy of the image retrieval result are further improved, so that the determination efficiency of the face images to be compared is further improved. Determining the image similarity between the face image to be identified and each sub-category face image in the image retrieval result; and screening the face images to be compared from the face images of each sub-category in the image retrieval result according to the similarity of each image. The sub-category face images in the image retrieval result are screened, and the number of face images to be compared is further reduced, so that the face recognition efficiency and the payment efficiency of face recognition payment are further improved.
Example III
The present embodiment provides a face recognition device, referring to fig. 5, which specifically includes:
the target face category determining module 510 is configured to obtain face feature information of a face image to be identified, and determine, based on the face feature information, a target face category to which the face image to be identified belongs by using a preset face classification model;
the face image to be compared determining module 520 is configured to determine a face image to be compared from a face classification database according to the target face class;
the face recognition result obtaining module 530 is configured to obtain a face recognition result by comparing the face image to be recognized with the face image to be compared.
Optionally, the face image to be compared determining module 520 includes:
the sub-category face image determining sub-module is used for matching the target face category with each face category in the face classification database, determining the matched face category, and determining the sub-category face image from the face classification database according to the matched face category;
the face image to be compared determining sub-module is used for carrying out image retrieval on the sub-category face images according to the face images to be identified and determining the face images to be compared from the sub-category face images according to the image retrieval result.
Optionally, the face image to be compared determining submodule is specifically configured to:
determining the image similarity between the face image to be identified and each sub-category face image in the image retrieval result;
and screening the face images to be compared from the face images of each sub-category in the image retrieval result according to the similarity of each image.
Further, the face image to be compared determining sub-module is further specifically configured to:
and determining the image similarity between the face image to be identified and the face image of the corresponding sub-category according to the query times of the face image of each sub-category in the image retrieval result, wherein the query times are the times of the face image of the sub-category in the image retrieval process.
Optionally, the sub-category face image determining submodule is specifically configured to:
extracting features of the face image to be identified, and obtaining a dimension reduction feature matrix of the face image to be identified based on the feature extraction result;
determining the vector similarity between each dimension-reduction feature vector and at least one preset retrieval feature vector in the dimension-reduction feature matrix;
determining a retrieval tag sequence of each dimension-reduction feature vector according to the similarity of each vector, a preset similarity threshold value and a preset retrieval tag corresponding to each preset retrieval feature vector;
And determining an image retrieval result according to the retrieval tag sequences of the dimensions of the face images to be identified and the retrieval tag sequences of the dimensions of the face images of each sub-category.
Further, the sub-category face image determination submodule is specifically configured to:
according to the query rule with the same preset retrieval tag in the same vector dimension, taking each dimension retrieval tag sequence of the face image to be recognized as a query standard, carrying out image query on each sub-category face image according to each dimension retrieval tag sequence of each sub-category face image, and determining each queried sub-category face image as an image retrieval result.
Optionally, on the basis of the above device, the device further includes a face classification database construction module, configured to pre-construct a face classification database by:
inputting each face image in a face database into a preset face classification model, and determining the face category to which each face image belongs;
and determining face images contained in the corresponding face categories from the face images belonging to each face category according to the number of the face images belonging to each face category and the number of the face images of a preset database, and constructing a face classification database, wherein the number of the face images of the preset database is the number of the face images at the falling inflection point of the face recognition rate.
Optionally, the preset face classification model is obtained by training at least one supervised classification machine learning model in advance.
According to the face recognition device provided by the embodiment of the invention, the local face database is divided into a plurality of different face categories, and face recognition is carried out in a certain face category, so that the problem of reduced face recognition efficiency caused by the fact that the total number of face images in the local face database exceeds the number of face images when the face recognition efficiency begins to be reduced is avoided, the total number of face images in the local face database can be increased, the face recognition efficiency is improved, and the applicability of face recognition payment to an actual payment scene is improved on the basis of face recognition payment convenience.
The face recognition device provided by the embodiment of the invention can execute the face recognition method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
It should be noted that, in the embodiment of the face recognition device, each unit and module included are only divided according to the functional logic, but not limited to the above division, so long as the corresponding functions can be implemented; in addition, the specific names of the functional units are also only for distinguishing from each other, and are not used to limit the protection scope of the present invention.
Example IV
Referring to fig. 6, the present embodiment provides an electronic device 600, which includes: one or more processors 620; the storage device 610 is configured to store one or more programs, where the one or more programs are executed by the one or more processors 620, so that the one or more processors 620 implement the face recognition method provided by the embodiment of the present invention, and the method includes:
acquiring face characteristic information of a face image to be identified, and determining a target face category to which the face image to be identified belongs by using a preset face classification model based on the face characteristic information;
determining face images to be compared from a face classification database according to the target face types;
and comparing the face image to be identified with the face image to be compared to obtain a face identification result.
Of course, those skilled in the art will appreciate that the processor 620 may also implement the technical solution of the face recognition method provided in any embodiment of the present invention.
The electronic device 600 shown in fig. 6 is merely an example, and should not be construed as limiting the functionality and scope of use of embodiments of the present invention.
As shown in fig. 6, the electronic device 600 is in the form of a general purpose computing device. Components of electronic device 600 may include, but are not limited to: one or more processors 620, a memory device 610, and a bus 650 that connects the different system components (including the memory device 610 and the processor 620).
Bus 650 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, and a local bus using any of a variety of bus architectures. By way of example, and not limitation, such architectures include Industry Standard Architecture (ISA) bus, micro channel architecture (MAC) bus, enhanced ISA bus, video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.
Electronic device 600 typically includes a variety of computer system readable media. Such media can be any available media that is accessible by electronic device 600 and includes both volatile and nonvolatile media, removable and non-removable media.
The storage 610 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM) 611 and/or cache memory 612. The electronic device 600 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 613 may be used to read from or write to non-removable, nonvolatile magnetic media (not shown in FIG. 6, commonly referred to as a "hard disk drive"). Although not shown in fig. 6, a magnetic disk drive for reading from and writing to a removable non-volatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading from or writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In such cases, each drive may be coupled to bus 650 via one or more data medium interfaces. Storage device 610 may include at least one program product having a set (e.g., at least one) of program modules configured to carry out the functions of embodiments of the invention.
A program/utility 614 having a set (at least one) of program modules 615 may be stored, for example, in the storage 610, such program modules 615 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each or some combination of which may include an implementation of a network environment. Program modules 615 generally perform the functions and/or methods of any of the embodiments described herein.
The electronic device 600 may also communicate with one or more external devices 660 (e.g., keyboard, pointing device, display 670, etc.), one or more devices that enable a user to interact with the electronic device 600, and/or any devices (e.g., network card, modem, etc.) that enable the electronic device 600 to communicate with one or more other computing devices. Such communication may occur through an input/output (I/O) interface 630. Also, the electronic device 600 may communicate with one or more networks such as a Local Area Network (LAN), a Wide Area Network (WAN) and/or a public network, such as the Internet, through the network adapter 640. As shown in fig. 6, network adapter 640 communicates with other modules of electronic device 600 over bus 650. It should be appreciated that although not shown, other hardware and/or software modules may be used in connection with electronic device 600, including, but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, data backup storage systems, and the like.
Example five
The present embodiment provides a storage medium containing computer executable instructions which, when executed by a computer processor, are configured to perform a face recognition method comprising:
acquiring face characteristic information of a face image to be identified, and determining a target face category to which the face image to be identified belongs by using a preset face classification model based on the face characteristic information;
determining face images to be compared from a face classification database according to the target face types;
and comparing the face image to be identified with the face image to be compared to obtain a face identification result.
Of course, the storage medium containing the computer executable instructions provided in the embodiments of the present invention is not limited to the above method operations, and may also perform the related operations in the face recognition method provided in any embodiment of the present invention.
The computer storage media of embodiments of the invention may take the form of any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
The computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, either in baseband or as part of a carrier wave. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).
Note that the above is only a preferred embodiment of the present invention and the technical principle applied. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, while the invention has been described in connection with the above embodiments, the invention is not limited to the embodiments, but may be embodied in many other equivalent forms without departing from the spirit or scope of the invention, which is set forth in the following claims.
Claims (10)
1. A face recognition method, comprising:
acquiring face characteristic information of a face image to be identified, and determining a target face category to which the face image to be identified belongs by using a preset face classification model based on the face characteristic information;
determining face images to be compared from a face classification database according to the target face types;
the face image to be identified is compared with the face image to be compared, so that a face identification result is obtained;
the step of determining the face image to be compared from the face classification database according to the target face category comprises the following steps:
Matching the target face category with each face category in the face classification database, determining a matched face category, and determining a sub-category face image from the face classification database according to the matched face category; wherein the sub-category face image refers to a face image in a face category; the face category is a category determined by classifying the types of faces of users;
and carrying out image retrieval on the sub-category face images according to the face images to be identified, and determining the face images to be compared from the sub-category face images according to an image retrieval result.
2. The method of claim 1, wherein determining the face image to be compared from the sub-category face images based on an image retrieval result comprises:
determining the image similarity between the face image to be identified and each sub-category face image in the image retrieval result;
and screening the face images to be compared from the face images of each sub-category in the image retrieval result according to the similarity of each image.
3. The method of claim 2, wherein determining the image similarity between the face image to be identified and each sub-category face image in the image retrieval result comprises:
And determining the image similarity between the face image to be identified and the face image of the corresponding sub-category according to the query times of the face image of each sub-category in the image retrieval result, wherein the query times are the times of the face image of the sub-category being queried in the image retrieval process.
4. The method of claim 1, wherein image retrieval of the sub-category face images from the face image to be identified comprises:
extracting features of the face image to be identified, and obtaining a dimension reduction feature matrix of the face image to be identified based on feature extraction results;
determining the vector similarity between each dimension-reduction feature vector and at least one preset retrieval feature vector in the dimension-reduction feature matrix;
determining a retrieval tag sequence of the dimension-reduction feature vector in each dimension according to the vector similarity, a preset similarity threshold and a preset retrieval tag corresponding to each preset retrieval feature vector;
and determining the image retrieval result according to the retrieval tag sequences of the face images to be identified and the retrieval tag sequences of the face images of each sub-category.
5. The method of claim 4, wherein determining the image search result based on the sequence of search tags for the dimensions of the face image to be identified and the sequence of search tags for the dimensions of each of the sub-category face images comprises:
according to the query rule with the same preset retrieval tag in the same vector dimension, taking each dimension retrieval tag sequence of the face image to be identified as a query standard, carrying out image query on each sub-category face image according to each dimension retrieval tag sequence of each sub-category face image, and determining each queried sub-category face image as the image retrieval result.
6. The method of claim 1, wherein the face classification database is pre-built by:
inputting each face image in a face database into the preset face classification model, and determining the face category to which each face image belongs;
and determining the face images contained in the corresponding face categories from the face images belonging to each face category according to the number of the face images belonging to each face category and the number of the face images of a preset database, and constructing the face classification database, wherein the number of the face images of the preset database is the number of the face images at the descending inflection point of the face recognition rate.
7. The method according to any one of claims 1-6, wherein the pre-set face classification model is obtained by training at least one supervised classification machine learning model in advance.
8. A face recognition device, comprising:
the target face type determining module is used for acquiring face characteristic information of a face image to be recognized and determining a target face type of the face image to be recognized by utilizing a preset face classification model based on the face characteristic information;
the face image to be compared determining module is used for determining face images to be compared from a face classification database according to the target face category;
the face recognition result obtaining module is used for obtaining a face recognition result by comparing the face image to be recognized with the face image to be compared;
the face image to be compared determining module comprises:
the sub-category face image determining sub-module is used for matching the target face category with each face category in the face classification database, determining the matched face category, and determining the sub-category face image from the face classification database according to the matched face category; wherein the sub-category face image refers to a face image in a face category; the face category is a category determined by classifying the types of faces of users;
The face image to be compared determining sub-module is used for carrying out image retrieval on the sub-category face images according to the face images to be identified and determining the face images to be compared from the sub-category face images according to the image retrieval result.
9. An electronic device, the electronic device comprising:
one or more processors;
storage means for storing one or more programs,
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the face recognition method of any one of claims 1-7.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements a face recognition method according to any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010112341.3A CN111325156B (en) | 2020-02-24 | 2020-02-24 | Face recognition method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010112341.3A CN111325156B (en) | 2020-02-24 | 2020-02-24 | Face recognition method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111325156A CN111325156A (en) | 2020-06-23 |
| CN111325156B true CN111325156B (en) | 2023-08-11 |
Family
ID=71172878
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010112341.3A Active CN111325156B (en) | 2020-02-24 | 2020-02-24 | Face recognition method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111325156B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111797746A (en) * | 2020-06-28 | 2020-10-20 | 北京小米松果电子有限公司 | Face recognition method and device and computer readable storage medium |
| CN112069875A (en) * | 2020-07-17 | 2020-12-11 | 北京百度网讯科技有限公司 | Face image classification method and device, electronic equipment and storage medium |
| CN112015933A (en) * | 2020-08-06 | 2020-12-01 | 北京奇艺世纪科技有限公司 | Image retrieval method, device and system, electronic equipment and storage medium |
| CN113065530B (en) * | 2021-05-12 | 2023-05-30 | 曼德电子电器有限公司 | Face recognition method and device, medium and equipment |
| CN113689613A (en) * | 2021-08-25 | 2021-11-23 | 爱康网健康科技(北京)有限公司 | Access control system, access control method, and storage medium |
| CN114612983A (en) * | 2022-03-15 | 2022-06-10 | 北京拙河科技有限公司 | Face recognition method and system based on elastic graph and SVM |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096517A (en) * | 2016-06-01 | 2016-11-09 | 北京联合大学 | A kind of face identification method based on low-rank matrix Yu eigenface |
| CN106202256A (en) * | 2016-06-29 | 2016-12-07 | 西安电子科技大学 | Propagate based on semanteme and mix the Web graph of multi-instance learning as search method |
| CN106250821A (en) * | 2016-07-20 | 2016-12-21 | 南京邮电大学 | The face identification method that a kind of cluster is classified again |
| CN106503687A (en) * | 2016-11-09 | 2017-03-15 | 合肥工业大学 | The monitor video system for identifying figures of fusion face multi-angle feature and its method |
| CN109241868A (en) * | 2018-08-16 | 2019-01-18 | 平安科技（深圳）有限公司 | Face identification method, device, computer equipment and storage medium |
| CN109426781A (en) * | 2017-08-29 | 2019-03-05 | 阿里巴巴集团控股有限公司 | Construction method, face identification method, device and the equipment of face recognition database |
| WO2019071664A1 (en) * | 2017-10-09 | 2019-04-18 | 平安科技（深圳）有限公司 | Human face recognition method and apparatus combined with depth information, and storage medium |
| CN109919093A (en) * | 2019-03-07 | 2019-06-21 | 苏州科达科技股份有限公司 | A kind of face identification method, device, equipment and readable storage medium storing program for executing |
| WO2019119505A1 (en) * | 2017-12-18 | 2019-06-27 | 深圳云天励飞技术有限公司 | Face recognition method and device, computer device and storage medium |
| CN110110593A (en) * | 2019-03-27 | 2019-08-09 | 广州杰赛科技股份有限公司 | Face Work attendance method, device, equipment and storage medium based on self study |
| CN110298249A (en) * | 2019-05-29 | 2019-10-01 | 平安科技（深圳）有限公司 | Face identification method, device, terminal and storage medium |
| CN110458154A (en) * | 2019-09-12 | 2019-11-15 | 腾讯科技（深圳）有限公司 | Face identification method, device and computer readable storage medium |
| WO2020015075A1 (en) * | 2018-07-18 | 2020-01-23 | 平安科技（深圳）有限公司 | Facial image comparison method and apparatus, computer device, and storage medium |
| CN110826525A (en) * | 2019-11-18 | 2020-02-21 | 天津高创安邦技术有限公司 | Face recognition method and system |
-
2020
- 2020-02-24 CN CN202010112341.3A patent/CN111325156B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096517A (en) * | 2016-06-01 | 2016-11-09 | 北京联合大学 | A kind of face identification method based on low-rank matrix Yu eigenface |
| CN106202256A (en) * | 2016-06-29 | 2016-12-07 | 西安电子科技大学 | Propagate based on semanteme and mix the Web graph of multi-instance learning as search method |
| CN106250821A (en) * | 2016-07-20 | 2016-12-21 | 南京邮电大学 | The face identification method that a kind of cluster is classified again |
| CN106503687A (en) * | 2016-11-09 | 2017-03-15 | 合肥工业大学 | The monitor video system for identifying figures of fusion face multi-angle feature and its method |
| CN109426781A (en) * | 2017-08-29 | 2019-03-05 | 阿里巴巴集团控股有限公司 | Construction method, face identification method, device and the equipment of face recognition database |
| WO2019071664A1 (en) * | 2017-10-09 | 2019-04-18 | 平安科技（深圳）有限公司 | Human face recognition method and apparatus combined with depth information, and storage medium |
| WO2019119505A1 (en) * | 2017-12-18 | 2019-06-27 | 深圳云天励飞技术有限公司 | Face recognition method and device, computer device and storage medium |
| WO2020015075A1 (en) * | 2018-07-18 | 2020-01-23 | 平安科技（深圳）有限公司 | Facial image comparison method and apparatus, computer device, and storage medium |
| CN109241868A (en) * | 2018-08-16 | 2019-01-18 | 平安科技（深圳）有限公司 | Face identification method, device, computer equipment and storage medium |
| CN109919093A (en) * | 2019-03-07 | 2019-06-21 | 苏州科达科技股份有限公司 | A kind of face identification method, device, equipment and readable storage medium storing program for executing |
| CN110110593A (en) * | 2019-03-27 | 2019-08-09 | 广州杰赛科技股份有限公司 | Face Work attendance method, device, equipment and storage medium based on self study |
| CN110298249A (en) * | 2019-05-29 | 2019-10-01 | 平安科技（深圳）有限公司 | Face identification method, device, terminal and storage medium |
| CN110458154A (en) * | 2019-09-12 | 2019-11-15 | 腾讯科技（深圳）有限公司 | Face identification method, device and computer readable storage medium |
| CN110826525A (en) * | 2019-11-18 | 2020-02-21 | 天津高创安邦技术有限公司 | Face recognition method and system |
Non-Patent Citations (1)
| Title |
|---|
| 基于分块的有遮挡人脸识别算法;周孝佳 等;计算机应用与软件(第02期);183-187 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111325156A (en) | 2020-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111325156B (en) | Face recognition method, device, equipment and storage medium | |
| US11514698B2 (en) | Intelligent extraction of information from a document | |
| WO2021072885A1 (en) | Method and apparatus for recognizing text, device and storage medium | |
| JP5211050B2 (en) | 2-step text recognition | |
| US20110295778A1 (en) | Information processing apparatus, information processing method, and program | |
| CN110941951A (en) | Text similarity calculation method, text similarity calculation device, text similarity calculation medium and electronic equipment | |
| CN114218945A (en) | Entity identification method, device, server and storage medium | |
| US20210157833A1 (en) | Visual image search using text-based search engines | |
| CN113486178B (en) | Text recognition model training method, text recognition method, device and medium | |
| US11625630B2 (en) | Identifying intent in dialog data through variant assessment | |
| WO2022007596A1 (en) | Image retrieval system, method and apparatus | |
| CN112579781B (en) | Text classification method, device, electronic equipment and medium | |
| CN116343233B (en) | Text recognition method and training method and device of text recognition model | |
| JP2001337993A (en) | Retrieval device and method for retrieving information by use of character recognition result | |
| CN115953123A (en) | Method, device and equipment for generating robot automation flow and storage medium | |
| CN111444319B (en) | Text matching method and device and electronic equipment | |
| CN115080745A (en) | Multi-scene text classification method, device, equipment and medium based on artificial intelligence | |
| WO2021159668A1 (en) | Robot dialogue method and apparatus, computer device and storage medium | |
| CN112668341A (en) | Text regularization method, device and equipment and readable storage medium | |
| CN112149389A (en) | Resume information structured processing method and device, computer equipment and storage medium | |
| Lakshmi | An efficient telugu word image retrieval system using deep cluster | |
| CN111783869A (en) | Training data screening method and device, electronic equipment and storage medium | |
| CN111382247A (en) | Content pushing optimization method, content pushing optimization device and electronic equipment | |
| CN117235137B (en) | Professional information query method and device based on vector database | |
| CN113177479B (en) | Image classification method, device, electronic equipment and storage medium |
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 |