Sample processing method and device, equipment and storage medium

Abstract

The invention provides a sample processing method and device, equipment and a storage medium, and relates to the technical field of data processing, in particular to the technical fields of artificial intelligence, big data and the like. According to the specific implementation scheme, a plurality of historical living body images are acquired; the historical living body image represents biological characteristic information of the target object; performing clustering processing on the plurality of historical living body images based on image features with biological feature information corresponding to the historical living body images to obtain a plurality of first clusters; obtaining a risk level of each first cluster based on the attribute feature information of the target object; and at least utilizing the risk level of each first cluster to construct a target sample data set.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to the technical fields of artificial intelligence, big data, etc. Background Art

[0002] In more and more scenarios, image-based liveness detection is often used to assess the risk of abnormal conditions of target objects in images, and the accuracy of risk assessment depends on the quality and diversity of sample data. However, existing data acquisition methods have problems such as high cost, low efficiency, and insufficient data diversity, which results in the inability to ensure data quality and low utilization. Summary of the invention

[0003] The present disclosure provides a sample processing method, apparatus, device and storage medium.

[0004] According to one aspect of the present disclosure, there is provided a sample processing method, comprising:

[0005] Acquire multiple historical living images; the historical living images represent the biometric information of the target object;

[0006] Based on image features with biometric information corresponding to the historical living body images, clustering the plurality of historical living body images to obtain a plurality of first clusters;

[0007] Based on the attribute characteristic information of the target object, the risk level of each first cluster is obtained;

[0008] At least the risk level of each first cluster is used to construct a target sample data set.

[0009] According to another aspect of the present disclosure, there is provided a sample processing device, comprising:

[0010] An acquisition unit, used for acquiring a plurality of historical living body images; the historical living body images represent the biometric information of the target object;

[0011] A processing unit is used to cluster the multiple historical living images based on the image features with biometric information corresponding to the historical living images to obtain multiple first clusters; obtain the risk level of each first cluster based on the attribute feature information of the target object; and construct a target sample data set using at least the risk level of each first cluster.

[0012] According to another aspect of the present disclosure, there is provided an electronic device, comprising:

[0013] at least one processor; and

[0014] a memory communicatively connected to the at least one processor; wherein,

[0015] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute any method in the embodiments of the present disclosure.

[0016] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute any method according to the embodiments of the present disclosure.

[0017] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which implements any method according to the embodiments of the present disclosure when executed by a processor.

[0018] In this way, the disclosed scheme can utilize the image features with biometric information corresponding to the historical living images to cluster the multiple historical living images, and grade the clustering results according to the attribute feature information of the target object, and then obtain the risk level of each first cluster. In this way, a large amount of high-quality sample data can be batch-mined using the determined first clusters with risk levels. The process is efficient and can effectively reduce the cost of data mining, provide sufficient training samples for subsequent model training and optimization, and at the same time, lay the foundation for improving the accuracy of living anomaly detection.

[0019] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0021] Figure 1 This is a schematic flow chart of a sample processing method according to an embodiment of the present application. Figure 1 ;

[0022] Figure 2 This is a schematic flow chart of a sample processing method according to an embodiment of the present application. Figure 2 ;

[0023] Figure 3 This is a schematic flow chart of a sample processing method according to an embodiment of the present application. Figure 3 ;

[0024] Figure 4 is a schematic structural diagram of a sample processing device according to an embodiment of the present application;

[0025] Figure 5 is a block diagram of an electronic device for implementing the sample processing method according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0026] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0027] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The term "at least one" in this article means any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C. The terms "first" and "second" in this article refer to multiple similar technical terms and distinguish them. They do not mean to limit the order or to limit them to only two. For example, the first feature and the second feature refer to two types / two features. The first feature can be one or more, and the second feature can also be one or more.

[0028] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the specific embodiments below. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present disclosure.

[0029] The disclosed solution provides a sample processing method, so that sufficient and high-quality sample data can be obtained for model training, and the obtained sample data is greatly improved in diversity and richness, which is of great significance for improving the accuracy of model inference results (the inference results can be used to predict whether a living body in an image is in an abnormal state).

[0030] Specifically, Figure 1 This is a schematic flow chart of a sample processing method according to an embodiment of the present application. Figure 1 The method may be optionally applied to electronic devices, such as personal computers, servers, server clusters, and other electronic devices.

[0031] Further, the method includes at least part of the following contents. Figure 1 As shown, including:

[0032] Step S101: Acquire multiple historical living body images.

[0033] Here, the historical living body image can represent the biometric information of the target object. For example, in one example, the historical living body image can specifically be an image containing facial features of the target object, such as a face image.

[0034] Furthermore, in another example, the historical living image may be an image that is generated using computer graphics technology (such as a generative adversarial network (GAN)) and is capable of representing the biometric information (such as facial features) of the target object; or, it may be an image that is obtained by enhancing the living feature information using data enhancement technology on the basis of the original living image.

[0035] It should be noted that the disclosed solution does not limit the method of acquiring historical living body images.

[0036] Step S102: performing clustering processing on the plurality of historical living body images based on image features with biometric feature information corresponding to the historical living body images to obtain a plurality of first clusters.

[0037] It should be noted that the "biometric information" in the image features may be feature information that can distinguish to the greatest extent whether the target object is abnormal, for example, living feature information that can distinguish to the greatest extent whether the target object is a living body, so as to facilitate the subsequent extraction of the required sample data to the greatest extent.

[0038] Furthermore, the clustering method in this example may be a density-based spatial clustering (Density-Based Spatial Clustering of Applications with Noise, DBSCAN) algorithm, or may be other clustering algorithms, and the present disclosure does not impose any specific limitation on this.

[0039] Step S103: obtaining the risk level of each first cluster based on the attribute feature information of the target object.

[0040] Here, in one example, the risk level of the first cluster may refer to the risk level of the “biometric information” contained in the first cluster, and further, may characterize the risk level of the abnormality of the target object corresponding to the first cluster.

[0041] For example, for the first cluster, the risk level of the first cluster can be determined based on the attribute feature information of each target object corresponding to the first cluster. In other words, in this example, after obtaining multiple first clusters, first, the attribute feature information of the target object corresponding to each historical living body image in the current first cluster is determined; second, the attribute feature information of the target object corresponding to each historical living body image in the current first cluster is used to determine the biometric information contained in the first cluster for risk assessment to obtain the risk level of the current first cluster; finally, the risk level of each first cluster is determined. In this way, a basis is provided for the subsequent selection of sample data that meets the requirements.

[0042] Here, in one example, the attribute feature information may specifically include: account information, security information, etc. In this way, the attribute feature information is used to obtain the risk level of each first cluster, thereby providing strong support for the subsequent mining of positive and negative sample data.

[0043] Step S104: construct a target sample data set using at least the risk levels of each first cluster.

[0044] That is to say, in the disclosed scheme, first, multiple historical living images are clustered according to image features carrying biometric information corresponding to the historical living images, so as to divide historical living images with similar biometric information into the same cluster, thereby obtaining multiple first clusters; secondly, the risk level of each first cluster is determined according to the attribute feature information of each target object corresponding to the first cluster; finally, a high-quality target sample data set is constructed according to the risk level of each first cluster.

[0045] In this way, the disclosed scheme can utilize the image features with biometric information corresponding to the historical living images to cluster the multiple historical living images, and grade the clustering results according to the attribute feature information of the target object, and then obtain the risk level of each first cluster. In this way, a large amount of high-quality sample data can be batch-mined using the determined first clusters with risk levels. The process is efficient and can effectively reduce the cost of data mining, provide sufficient training samples for subsequent model training and optimization, and at the same time, lay the foundation for improving the accuracy of living anomaly detection.

[0046] Furthermore, since the disclosed scheme first clusters the multiple historical living images into multiple first clusters according to the image features with biometric information corresponding to the historical living images, and then determines the risk level of each first cluster, it is also possible to effectively monitor the first clusters with high risk levels with the help of the risk levels of each first cluster, thereby reducing the workload and complexity of risk monitoring, and helping to manually discover and deal with potential problems in a timely manner (for example, non-living account management operations, etc.), thereby effectively ensuring the normal operation of the entire system.

[0047] In a specific example, the image features with biometric information corresponding to the historical living images can be obtained in the following manner; for example, before the above-mentioned step S102, the historical living images can also be input into the pre-trained image feature extraction model to obtain the image features with biometric information corresponding to the historical living images, for example, to obtain the vectorized features with biometric information corresponding to the historical living images (for example, which can be represented by Embedding). In this way, a basis is provided for the subsequent clustering of multiple historical living images, thereby laying a foundation for constructing a high-quality target sample data set.

[0048] In a specific example, the target sample data set may be constructed in the following three ways; specifically, the above-mentioned construction of the target sample data set (for example, step S104) by at least using the risk level of each first cluster may specifically include:

[0049] Method 1: Determine the first cluster whose risk level meets the preset conditions, and construct positive sample data based on the image features with biometric information corresponding to the first cluster that meets the preset conditions to obtain the target sample data set.

[0050] Here, the preset condition may specifically be that the risk level is less than or equal to a preset risk threshold. In practical applications, it may be set according to actual needs, and the present disclosure does not limit this.

[0051] For example, in one example, based on the risk level of each first cluster, at least one first cluster whose risk level is less than or equal to a preset risk threshold is determined from multiple first clusters, and then the image features with biometric information corresponding to each of the at least one first cluster determined are used as positive sample data (i.e., sample data with no risk or low risk) to obtain a target sample data set. Alternatively, in another example, based on the image features with biometric information corresponding to each of the at least one first cluster determined, image features with a feature concentration lower than a threshold are determined, and the image features with a concentration lower than the threshold are used as positive sample data to obtain the target sample data.

[0052] It should be pointed out that the image features with biometric information corresponding to the first cluster that meets the preset condition specifically refer to the image features with biometric information corresponding to each historical living body image in the first cluster that meets the preset condition.

[0053] Method 2: Determine the first cluster whose risk level does not meet the preset conditions, and construct negative sample data based on the image features with biometric information corresponding to the first cluster that does not meet the preset conditions to obtain the target sample data set.

[0054] For example, in one example, based on the risk level of each first cluster, at least one first cluster whose risk level is greater than a preset risk threshold is determined from multiple first clusters, and then the image features with biometric information corresponding to each of the at least one first cluster determined are used as negative sample data (i.e., high-risk sample data) to obtain a target sample data set. Alternatively, in another example, based on the image features with biometric information corresponding to each of the at least one first cluster determined, image features with a feature concentration higher than a threshold are determined, and the image features with a concentration higher than the threshold are used as negative sample data to obtain the target sample data.

[0055] It should be pointed out that the image features with biometric information corresponding to the first cluster that does not meet the preset condition specifically refer to the image features with biometric information corresponding to each historical living body image in the first cluster that does not meet the preset condition.

[0056] Method 3: Determine whether the risk level of the first cluster meets the preset conditions. If so, construct positive sample data based on the image features with biometric information corresponding to the first cluster that meets the preset conditions; otherwise, construct negative sample data based on the image features with biometric information corresponding to the first cluster that does not meet the preset conditions, and then obtain the target sample data set based on the obtained positive sample data and negative sample data.

[0057] For example, after obtaining the risk level of each first cluster, determine whether the risk level of the current first cluster meets the preset conditions, such as whether it is less than or equal to the preset risk threshold; if the risk level of the current first cluster is less than or equal to the preset risk threshold, the image features with biometric information corresponding to the current first cluster are used as positive sample data, otherwise as negative sample data; in this way, traverse all first clusters to obtain all positive sample data and all negative sample data, thereby obtaining the target sample data set. In this way, high-quality sample data can be mined in batches to provide data support for subsequent model training and optimization.

[0058] In this way, the disclosed scheme can construct positive sample data and / or negative sample data according to the risk level of each first cluster and the image features with biometric information corresponding to each first cluster to obtain the target sample data set. In this way, high-quality sample data can be obtained at low cost, providing more and high-quality training samples for subsequent model training and optimization, thereby laying the foundation for improving the accuracy of living abnormality detection.

[0059] Figure 2 This is a schematic flow chart of a sample processing method according to an embodiment of the present application. Figure 2The method can optionally be applied to electronic devices, such as personal computers, servers, server clusters, and other electronic devices. It is understandable that the above Figure 1 The relevant contents of the method shown can also be applied to this example, and the relevant contents will not be described in detail in this example.

[0060] Further, the method includes at least part of the following contents. Specifically, Figure 2 As shown, including:

[0061] Step S201: Acquire multiple historical living body images.

[0062] Here, historical living images can characterize the biometric information of the target object.

[0063] It should be noted that the relevant examples of historical living body images can be found in the above description and will not be repeated here.

[0064] Step S202: performing clustering processing on the plurality of historical living body images based on image features with biometric feature information corresponding to the historical living body images to obtain a plurality of first clusters.

[0065] Here, the relevant content about the clustering method can refer to the above example, which will not be repeated here.

[0066] Step S203: obtaining the risk level of each first cluster based on the attribute feature information of the target object.

[0067] Here, the risk level of the first cluster can represent the risk level of the biometric information included in the first cluster.

[0068] Here, the attribute feature information of the target object can refer to the above examples, which will not be repeated here.

[0069] Step S204: construct a target sample data set using at least the risk levels of each first cluster.

[0070] Step S205: Acquire multiple historical feature data.

[0071] Here, the historical feature data represents the historical feature information of the target object. For example, in one example, the historical feature information may include historical operation information of the target object. For example, in one example, the historical operation information may specifically include information such as registration operation and verification operation of the target object.

[0072] Step S206: Based on the risk factors corresponding to the historical feature data and capable of characterizing the risk level of the historical feature information, cluster the plurality of historical feature data to obtain a second cluster for the risk factors. For example, multiple second clusters for the risk factors are obtained.

[0073] For example, for detecting whether a target object contained in an image is abnormal, such as whether it is a living body, the risk factor may be specifically the device information, network environment information, or geographic information of the target object. In practical applications, the risk factor may also be determined based on actual needs, and the disclosed solution does not impose specific restrictions on this.

[0074] Further, in one example, the risk factor may be obtained in the following manner; specifically, before clustering the plurality of historical feature data based on the risk factor that can characterize the risk level of the historical feature information corresponding to the historical feature data to obtain a second cluster for the risk factor (for example, before step S206), the method further includes:

[0075] Obtaining an initial factor having a clustering characteristic from the plurality of historical feature data;

[0076] Based on the initial factors having the aggregation characteristic, at least one risk factor capable of characterizing the risk level of the historical characteristic information is determined.

[0077] In practical applications, the operation data of the target object is usually presented in batches. These batch operation data may contain a lot of common information, such as the same geographical location, network environment, etc. At this time, these common information can be called initial factors, and the commonality of the initial factors is called aggregation characteristics; further, from the above initial factors with aggregation characteristics, risk factors that can reflect the risk level of historical characteristic information are selected. In this way, it is convenient to use risk factors to accurately identify the abnormal risk level of the target object corresponding to the historical characteristic data (for example, the risk level of not being a living body), thereby laying the foundation for subsequent targeted clustering and sample construction.

[0078] Step S207: Expanding the target sample data set based on the aggregation degree of the second cluster of risk factors.

[0079] For example, in one example, new sample data may be obtained based on the aggregation degree of the second cluster for the risk factor, and then the obtained new sample data may be added to the target sample data set obtained in step S204 to expand the target sample data set.

[0080] Furthermore, in one example, multiple risk factors can be selected. At this time, for each risk factor, steps S206 and S207 can be executed to obtain new sample data corresponding to each risk factor; further, the new sample data corresponding to each risk factor can be added to the target sample data set obtained in step S204 to expand the target sample data set. In this way, the data volume and richness of the target sample data set are further improved.

[0081] It should also be pointed out that in practical applications, the above example can also be to first use multiple historical feature data to obtain a target sample data set (for example, step S205 to step S207, hereinafter referred to as sample generation process 1), and then use multiple historical living images to obtain a new sample data set (that is, step S201 to step S204, hereinafter referred to as sample generation process 2), and use the new sample data set to expand the target sample data set; or, perform sample generation process 1 and sample generation process 2 at the same time to obtain two sets of sample data sets, and obtain the final target sample data set based on the two sets of sample data sets. The present disclosure does not specifically limit the execution order of the above two sample generation processes.

[0082] In this way, the disclosed scheme further expands the target sample data set by using a sample data set obtained by using a plurality of historical feature data on the basis of obtaining the target sample data set by using a plurality of historical living images; thus, the data volume and richness of the sample data set are improved, and at the same time, the data quality of the sample data is also improved, thereby providing a large number of training samples for subsequent model training and optimization, and further laying the foundation for improving the accuracy of living anomaly detection.

[0083] Further, in a specific example, the above-mentioned expanding the target sample data set based on the aggregation degree of the second cluster for the risk factor (for example, step S207) specifically includes:

[0084] Method 1: Determine a second cluster whose aggregation degree is less than a preset value, and obtain positive sample data based on the aggregation feature information corresponding to the risk factors in the second cluster whose aggregation degree is less than the preset value to expand the target sample data set.

[0085] For instance, in one example, based on the degree of clustering of each first cluster, at least one second cluster whose degree of clustering is less than a preset value is determined from multiple second clusters, and then the clustering feature information corresponding to the risk factors in each second cluster of the at least one determined second cluster is used as positive sample data (i.e., sample data with no risk or low risk) to expand the target sample data set.

[0086] Method 2: Determine a second cluster whose aggregation degree is greater than or equal to a preset value, and construct negative sample data based on the aggregation feature information corresponding to the risk factors in the second cluster greater than or equal to the preset value to expand the target sample data set.

[0087] For instance, in one example, based on the degree of clustering of each first cluster, at least one second cluster having a degree of clustering greater than or equal to a preset value is determined from multiple second clusters, and then the clustering feature information corresponding to the risk factors in each second cluster of the at least one determined second cluster is used as negative sample data (i.e., high-risk sample data) to expand the target sample data set.

[0088] Method 3: Determine whether the aggregation degree of the second cluster is less than a preset value. If so, construct positive sample data based on the aggregation feature information corresponding to the risk factors in the second cluster that are less than the preset value; otherwise, construct negative sample data based on the aggregation feature information corresponding to the risk factors in the second cluster that are greater than or equal to the preset value; and then obtain the target sample data set based on the obtained positive sample data and negative sample data.

[0089] Here, in this example, all second clusters are traversed to obtain all positive sample data and all negative sample data, thereby expanding the target sample data set based on all the positive sample data and all the negative sample data obtained. In this way, high-quality sample data is mined in batches to provide data support for subsequent model training and optimization.

[0090] In this way, the disclosed scheme can construct positive sample data and / or negative sample data according to the degree of clustering of each second cluster and the clustering feature information corresponding to the risk factors in each second cluster, so as to further expand the target sample data set. The process is simple and efficient, and high-quality sample data can be obtained at low cost. In this way, more high-quality training samples are provided for subsequent model training and optimization, thereby laying the foundation for improving the accuracy of living abnormality detection.

[0091] Figure 3 This is a schematic flow chart of a sample processing method according to an embodiment of the present application. Figure 3 The method can be optionally applied to electronic devices, such as personal computers, servers, server clusters, and other electronic devices. It can be understood that the above Figure 1 and Figure 2 The relevant contents of the method shown can also be applied to this example, and the relevant contents will not be described in detail in this example.

[0092] Further, the method includes at least part of the following contents. Figure 3 As shown, including:

[0093] Step S301: Acquire multiple historical living body images.

[0094] Here, historical living images can characterize the biometric information of the target object.

[0095] It should be noted that the relevant examples of historical living body images can be found in the above description and will not be repeated here.

[0096] Step S302: performing clustering processing on the plurality of historical living body images based on image features with biometric information corresponding to the historical living body images to obtain a plurality of first clusters.

[0097] Here, the relevant content about the clustering method can be referred to the above description, which will not be repeated here.

[0098] Step S303: obtaining the risk level of each first cluster based on the attribute feature information of the target object.

[0099] Here, the risk level of the first cluster can represent the risk level of the biometric information included in the first cluster.

[0100] Here, the relevant contents about the attribute feature information of the target object can be referred to the above description, which will not be repeated here.

[0101] Step S304: construct a target sample data set using at least the risk levels of each first cluster.

[0102] Step S305: Using the target sample data set, a preset detection model is trained to obtain a target detection model that can at least perform risk assessment based on living body images.

[0103] For example, in a specific example, the target sample data set and at least one of the following can be used to train the preset detection model: image features with biometric information corresponding to historical living images, clustered feature information corresponding to risk factors in historical feature data, and attribute feature information of target objects. This provides strong support for further improving the accuracy of model results.

[0104] Furthermore, in one example, the target inspection model is used to perform risk assessment on whether the target object in the input image is alive, for example, specifically used to detect whether the target object in the input image is alive.

[0105] Here, the preset detection model may be a gradient boosting decision tree (GBDT), or may be other integrated models, which is not limited in the present disclosure.

[0106] In this way, the disclosed solution can use the obtained target sample data set to train the preset detection model to obtain a target detection model. In this way, it can more accurately identify whether the target object in the image is abnormal, thereby providing strong support for effectively avoiding batch operations on non-living objects.

[0107] In addition, high-quality target sample data sets are conducive to improving the adaptability of the model in different scenarios, thereby improving the generalization ability of the model. In addition, using high-quality sample data to train the model can also improve the training efficiency of the model.

[0108] The disclosed solution also provides a sample processing device, such as Figure 4 As shown, including:

[0109] An acquisition unit 401 is used to acquire a plurality of historical living body images; the historical living body images represent the biometric information of the target object;

[0110] The processing unit 402 is used to perform clustering processing on the multiple historical living images based on the image features with biometric information corresponding to the historical living images to obtain multiple first clusters; obtain the risk level of each first cluster based on the attribute feature information of the target object; and construct a target sample data set using at least the risk level of each first cluster.

[0111] In a specific example of the disclosed solution, the processing unit is specifically used to:

[0112] Determine a first cluster whose risk level meets a preset condition, and construct positive sample data based on image features with biometric information corresponding to the first cluster that meets the preset condition to obtain a target sample data set;

[0113] and / or,

[0114] A first cluster whose risk level does not meet the preset condition is determined, and negative sample data is constructed based on image features with biometric information corresponding to the first cluster that does not meet the preset condition to obtain a target sample data set.

[0115] In a specific example of the disclosed solution, the acquisition unit is further used to: acquire a plurality of historical feature data; the historical feature data represents historical feature information of the target object;

[0116] The processing unit is also used to cluster the multiple historical feature data based on the risk factors corresponding to the historical feature data and capable of characterizing the risk level of the historical feature information to obtain a second cluster for the risk factors; and expand the target sample data set based on the degree of clustering of the second cluster for the risk factors.

[0117] In a specific example of the disclosed solution, the processing unit is further used for:

[0118] Obtaining an initial factor having a clustering characteristic from the plurality of historical feature data;

[0119] Based on the initial factors having the aggregation characteristic, at least one risk factor capable of characterizing the risk level of the historical characteristic information is determined.

[0120] In a specific example of the disclosed solution, the processing unit is specifically used to:

[0121] Determine a second cluster whose aggregation degree is less than a preset value, and construct positive sample data based on the aggregation feature information corresponding to the risk factors in the second cluster whose aggregation degree is less than the preset value to expand the target sample data set;

[0122] and / or,

[0123] A second cluster whose aggregation degree is greater than or equal to a preset value is determined, and negative sample data is constructed based on the aggregation feature information corresponding to the risk factors in the second cluster that are greater than or equal to the preset value to expand the target sample data set.

[0124] In a specific example of the disclosed solution, it also includes: a model training unit; wherein,

[0125] The model training unit is used to use the target sample data set to perform model training on a preset detection model to obtain a target detection model that can at least perform risk assessment based on living body images.

[0126] For the description of the specific functions and examples of each unit of the device in the embodiment of the present disclosure, reference can be made to the relevant description of the corresponding steps in the above method embodiment, which will not be repeated here.

[0127] In the technical solution disclosed herein, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0128] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0129] Figure 5 A schematic block diagram of an example electronic device 500 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0130] like Figure 5As shown, the device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0131] A number of components in the device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0132] The computing unit 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 501 performs the various methods and processes described above, such as the sample processing method. For example, in some embodiments, the sample processing method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 500 via ROM 502 and / or communication unit 509. When the computer program is loaded into RAM 503 and executed by the computing unit 501, one or more steps of the sample processing method described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to perform the sample processing method in any other appropriate manner (e.g., by means of firmware).

[0133] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0134] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0135] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, 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 disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0136] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0137] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0138] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0139] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0140] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A sample processing method, comprising: Acquire multiple historical in vivo images; Historical living images represent the biometric information of the target object; Based on image features with biometric information corresponding to the historical living body images, clustering the plurality of historical living body images to obtain a plurality of first clusters; Based on the attribute characteristic information of the target object, the risk level of each first cluster is obtained; At least the risk level of each first cluster is used to construct a target sample data set.

2. The method according to claim 1, wherein: The step of constructing a target sample data set by at least using the risk level of each first cluster includes: Determine a first cluster whose risk level meets a preset condition, and construct positive sample data based on image features with biometric information corresponding to the first cluster that meets the preset condition to obtain a target sample data set; and / or, A first cluster whose risk level does not meet the preset condition is determined, and negative sample data is constructed based on image features with biometric information corresponding to the first cluster that does not meet the preset condition to obtain a target sample data set.

3. The method according to claim 1 or 2, further comprising: Obtain multiple historical feature data; history Feature data represents the historical feature information of the target object; Based on the risk factors corresponding to the historical feature data and capable of characterizing the risk level of the historical feature information, clustering the plurality of historical feature data to obtain a second cluster for the risk factors; Based on the clustering degree of the second cluster for the risk factors, the target sample dataset is expanded.

4. The method according to claim 3, further comprising: Obtaining an initial factor having a clustering characteristic from the plurality of historical feature data; Based on the initial factors having the aggregation characteristic, at least one risk factor capable of characterizing the risk level of the historical characteristic information is determined.

5. The method according to claim 3, wherein: The step of expanding the target sample data set based on the aggregation degree of the second cluster of risk factors includes: Determine a second cluster whose aggregation degree is less than a preset value, and construct positive sample data based on the aggregation feature information corresponding to the risk factors in the second cluster whose aggregation degree is less than the preset value to expand the target sample data set; and / or, A second cluster whose aggregation degree is greater than or equal to a preset value is determined, and negative sample data is constructed based on the aggregation feature information corresponding to the risk factors in the second cluster that are greater than or equal to the preset value to expand the target sample data set.

6. The method according to any one of claims 1 to 4, further comprising: The target sample data set is used to perform model training on a preset detection model to obtain a target detection model that can at least perform risk assessment based on living body images.

7. A sample processing device, comprising: An acquisition unit, used for acquiring a plurality of historical living body images; Historical living images represent the biometric information of the target object; A processing unit, configured to perform clustering processing on the plurality of historical living body images based on image features with biometric information corresponding to the historical living body images, to obtain a plurality of first clusters; Based on the attribute characteristic information of the target object, the risk level of each first cluster is obtained; At least the risk level of each first cluster is used to construct a target sample data set.

8. The device according to claim 7, wherein: The processing unit is specifically used for: Determine a first cluster whose risk level meets a preset condition, and construct positive sample data based on image features with biometric information corresponding to the first cluster that meets the preset condition to obtain a target sample data set; and / or, A first cluster whose risk level does not meet the preset condition is determined, and negative sample data is constructed based on image features with biometric information corresponding to the first cluster that does not meet the preset condition to obtain a target sample data set.

9. The device according to claim 7 or 8, wherein: The acquisition unit is further used to acquire a plurality of historical feature data; the historical feature data represents the historical feature information of the target object; The processing unit is further configured to perform clustering processing on the plurality of historical feature data based on a risk factor corresponding to the historical feature data and capable of characterizing the risk level of the historical feature information, so as to obtain a second cluster for the risk factor; Based on the clustering degree of the second cluster for the risk factors, the target sample dataset is expanded.

10. The device according to claim 9, wherein: The processing unit is further used for: Obtaining an initial factor having a clustering characteristic from the plurality of historical feature data; Based on the initial factors having the aggregation characteristic, at least one risk factor capable of characterizing the risk level of the historical characteristic information is determined.

11. The device according to claim 9, wherein: The processing unit is specifically used for: Determine a second cluster whose aggregation degree is less than a preset value, and construct positive sample data based on the aggregation feature information corresponding to the risk factors in the second cluster whose aggregation degree is less than the preset value to expand the target sample data set; and / or, A second cluster whose aggregation degree is greater than or equal to a preset value is determined, and negative sample data is constructed based on the aggregation feature information corresponding to the risk factors in the second cluster that are greater than or equal to the preset value to expand the target sample data set.

12. The device according to any one of claims 7 to 10, further comprising: Model training unit; wherein, The model training unit is used to use the target sample data set to perform model training on a preset detection model to obtain a target detection model that can at least perform risk assessment based on living body images.

13. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-6.

15. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images