Method and apparatus for generating face image based on genetic rule

Abstract

The application relates to a method and device for generating a human face image based on genetic rules, which comprises the following steps: obtaining a target kinship image and a target description text, the target kinship image being an image of an object having a blood relationship with a target object; extracting a target kinship feature from the target kinship image; detecting whether a target description feature is extracted from the target description text, the target description feature being used for describing face feature information of the target object; when the target description feature is not extracted from the target description text, calling a face generation model to process the target kinship feature to obtain a human face image of the target object; when the target description feature is extracted from the target description text, detecting whether the target description feature comprises a genetic description feature and a face description feature to obtain a detection result, and calling the face generation model to process the target kinship feature and the feature in the detection result to obtain the human face image of the target object.

Description

Technical Field

[0001] This application relates to the field of genetic control technology, specifically to a method and apparatus for generating human face images based on genetic laws. Background Technology

[0002] In today's society, the rapid development of technology has brought about tremendous changes in various fields. Among them, image recognition and processing technology plays a crucial role in many scenarios, and its acquisition and utilization of facial images has had a profound impact on many aspects. However, in handling cases involving missing or maliciously transferred young children, unique challenges arise. In these cases, accurate photographs of the children are crucial for quickly locating them and confirming their identities. However, in reality, parents may be unable to provide clear and accurate photographs of their children for various reasons. Some families may have very few photos of their children due to limited living conditions. Therefore, how to obtain images of children that truly reflect their appearance under the predicament of lacking accurate photographs has become a critical issue that urgently needs to be addressed in the handling of such cases. Summary of the Invention

[0003] This application provides a method and apparatus for generating facial images based on genetic principles. This method can generate a facial image of a target person based on images of their relatives and a textual description of the target person. The generated facial image can highly realistically reflect the current facial features of a missing or maliciously transferred young person, providing crucial and powerful evidence for handling such cases.

[0004] A first aspect of this application provides a method for generating face images based on genetic laws, the method comprising:

[0005] Obtain the target kinship image and the target description text, wherein the target kinship image is an image of an object that is related to the target object by blood;

[0006] Extract target kinship features from the target kinship image, wherein the target kinship features are facial feature information of objects that are related to the target object by blood;

[0007] Detect whether target description features are extracted from the target description text, wherein the target description features are used to describe the facial feature information of the target object;

[0008] When the target description features are not extracted from the target description text, the face generation model is invoked to process the target kinship features to obtain the face image of the target object;

[0009] When extracting target description features from the target description text, it is detected whether the target description features include genetic description features and facial description features. The detection result is obtained, and a face generation model is called to process the target kinship features and the features in the detection result to obtain the face image of the target object.

[0010] Optionally, the face generation model includes a fusion module and a first generation module. The step of calling the face generation model to process the target kinship features and the features in the detection results to obtain the face image of the target object includes:

[0011] The fusion module is used to fuse the target kinship features and the features in the detection results to obtain fused features;

[0012] The fusion features are input into the first generation module so that the first generation module processes the fusion features to obtain the face image of the target object.

[0013] Optionally, the detection result includes genetic descriptive features and facial descriptive features, the face generation model includes a second generation module and an image modification model, and the step of calling the face generation model to process the target kinship features and the features in the detection result to obtain the face image of the target object includes:

[0014] Using the second generation module, the target kinship features and the genetic description features are processed to obtain a reference face image;

[0015] The image modification module is used to process the facial description features and the reference facial image to obtain the facial image of the target object.

[0016] Optionally, the method further includes:

[0017] The target description text is identified to obtain genetic description text and face description text;

[0018] The genetic description text is identified to obtain the genetic description features, and the facial description text is identified to obtain the facial description features.

[0019] Optionally, processing the target kinship features and the genetic description features to obtain the reference face image includes:

[0020] Obtain the first weight corresponding to the target kinship feature and the second weight corresponding to the genetic description feature;

[0021] Based on the first weight and the second weight, the target kinship feature and the genetic description feature are fused to obtain a first fused feature;

[0022] The first fused feature is processed to obtain a reference face image.

[0023] Optionally, the step of calling a face generation model to process the target kinship features to obtain the face image of the target object includes:

[0024] Detect whether the target kinship characteristic is associated with a genetic feature in the genetic feature database;

[0025] When the genetic feature library contains related genetic features of the target kinship feature, the face generation model is invoked to process the target kinship feature and the related genetic features to obtain the face image of the target object.

[0026] Optionally, processing the facial description features and the reference facial image to obtain the facial image of the target object includes:

[0027] Obtain the third weight corresponding to the face description feature and the fourth weight corresponding to the reference face image;

[0028] Based on the third weight and the fourth weight, the face description features and the reference face image are fused to obtain the second fused feature;

[0029] The second fusion feature is processed to obtain the face image of the target object.

[0030] In a second aspect, this application provides an apparatus for generating human face images based on genetic principles, comprising:

[0031] The acquisition unit is used to acquire a target kinship image and the target description text, wherein the target kinship image is an image of an object that is related to the target object by blood, and the target description text includes facial description information of the target object;

[0032] Extraction unit, used to extract target kinship features from the target kinship image;

[0033] The detection unit is used to detect whether target description features are extracted from the target description text;

[0034] The first calling unit is used to call a face generation model to process the target kinship features and obtain the face image of the target object when the target description features are not extracted from the target description text.

[0035] The second calling unit is used to detect whether the target description features include genetic description features and facial description features when extracting target description features from the target description text, obtain the detection result, call the face generation model, process the target kinship features and the features in the detection result, and obtain the face image of the target object.

[0036] Optionally, the face generation model includes a fusion module and a first generation module, and a second calling unit, used for:

[0037] The fusion module is used to fuse the target kinship features and the features in the detection results to obtain fused features;

[0038] The fusion features are input into the first generation module so that the first generation module processes the fusion features to obtain the face image of the target object.

[0039] Optionally, the detection result includes genetic descriptive features and facial descriptive features, the facial generation model includes a second generation module and an image modification model, and the second calling unit is used for:

[0040] Using the second generation module, the target kinship features and the genetic description features are processed to obtain a reference face image;

[0041] The image modification module is used to process the facial description features and the reference facial image to obtain the facial image of the target object.

[0042] Optionally, the device further includes an identification unit, the identification unit being used for:

[0043] The target description text is identified to obtain genetic description text and face description text;

[0044] The genetic description text is identified to obtain the genetic description features, and the facial description text is identified to obtain the facial description features.

[0045] Optionally, the second calling unit is used for:

[0046] Obtain the first weight corresponding to the target kinship feature and the second weight corresponding to the genetic description feature;

[0047] Based on the first weight and the second weight, the target kinship feature and the genetic description feature are fused to obtain a first fused feature;

[0048] The first fused feature is processed to obtain a reference face image.

[0049] Optionally, the first calling unit is used for:

[0050] Detect whether the target kinship characteristic is associated with a genetic feature in the genetic feature database;

[0051] When the genetic feature library contains related genetic features of the target kinship feature, the face generation model is invoked to process the target kinship feature and the related genetic features to obtain the face image of the target object.

[0052] Optionally, the second calling unit is used for:

[0053] Obtain the third weight corresponding to the face description feature and the fourth weight corresponding to the reference face image;

[0054] Based on the third weight and the fourth weight, the face description features and the reference face image are fused to obtain the second fused feature;

[0055] The second fusion feature is processed to obtain the face image of the target object.

[0056] A third aspect of this application provides a computer device, including: a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the above methods.

[0057] A fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps of the method as described in any of the above.

[0058] In this embodiment, a target kinship image and a target description text are obtained. The target kinship image is an image of an object related to the target by blood, and the target description text includes facial description information of the target object. Target kinship features are extracted from the target kinship image. It is detected whether target description features are extracted from the target description text. When no target description features are extracted from the target description text, a face generation model is invoked to process the target kinship features to obtain a face image of the target object. When target description features are extracted from the target description text, it is detected whether the target description features include genetic description features and facial description features. A detection result is obtained, and a face generation model is invoked to process the target kinship features and the features in the detection result to obtain a face image of the target object. This application can generate a face image of the target object based on an image of the target object's relatives, combined with a textual description of the target object. The generated face image can highly realistically reflect the current facial features of a missing or maliciously transferred young object, providing crucial and powerful evidence for handling such cases. Attached Figure Description

[0059] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0060] Figure 1 A flowchart illustrating a method for generating face images based on genetic principles, provided in one embodiment of this application;

[0061] Figure 2 A flowchart illustrating a face image generation method provided in one embodiment of this application;

[0062] Figure 3 A flowchart of a reference face image determination method provided in one embodiment of this application;

[0063] Figure 4 A flowchart illustrating a face image determination method provided in one embodiment of this application;

[0064] Figure 5 A bar chart showing the number of cases involving the malicious transfer of young victims solved in the middle of the period for the method of determining the target genetic characteristics of this application;

[0065] Figure 6 The number of times the ranking of a 2 million-repository library improved under different original library ranking intervals in the method for determining the target genetic characteristics of this application;

[0066] Figure 7 This is a schematic diagram of the structure of a face image generation device based on genetic laws provided in one embodiment of this application;

[0067] Figure 8 This is a schematic diagram of a computer device structure provided in one embodiment of this application. Detailed Implementation

[0068] In today's society, the rapid development of technology has brought about tremendous changes in various fields. Among them, image recognition and processing technology plays a crucial role in many scenarios, and its acquisition and utilization of facial images has had a profound impact on many aspects. However, in handling cases involving missing or maliciously transferred young children, unique challenges arise. In these cases, accurate photographs of the children are crucial for quickly locating them and confirming their identities. However, in reality, parents may be unable to provide clear and accurate photographs of their children for various reasons. Some families may have very few photos of their children due to limited living conditions. Therefore, how to obtain images of children that truly reflect their appearance under the predicament of lacking accurate photographs has become a critical issue that urgently needs to be addressed in the handling of such cases.

[0069] To address the aforementioned issues, this application provides a method for generating facial images based on genetic principles. First, images of individuals related to the target subject by blood are acquired, i.e., target kinship images. Simultaneously, target descriptive text containing facial description information of the target subject is also acquired. Then, target kinship features reflecting the kinship relationship are extracted from the target kinship images, and target descriptive features helpful in depicting the target subject's appearance are extracted from the target descriptive text. Finally, an advanced facial generation model is invoked, using these two types of features as input for processing, thereby generating a facial image of the target subject. The method provided in this application can fully utilize image and text information and efficiently process it using a model to generate a child's photograph. This photograph is of great significance in determining the child's identity, allowing relevant personnel to conduct further identity verification work, such as playing a crucial role in scenarios like finding missing young children and paternity testing.

[0070] The solutions in this application embodiment can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0071] To make the technical solutions and advantages of the embodiments of this application clearer, the exemplary embodiments of this application will be described in further detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not an exhaustive list of all embodiments. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.

[0072] Please see Figure 1 The following embodiments use a face generation system as the execution subject, applying the method provided in the embodiments of this application to the aforementioned face generation system. The method for generating face images based on genetic laws provided in the embodiments of this application includes the following steps 101-105:

[0073] Step 101: Obtain the target kinship image and the target description text.

[0074] The target kinship images are images of individuals related to the target subject by blood, typically the target subject's father or mother, but may also include images of siblings, grandparents, maternal grandparents, and other direct or collateral relatives. These kinship images carry diverse information about family genetic characteristics, helping to more comprehensively analyze the genetic traits the target subject may inherit. The target description text includes facial descriptions of the target subject, specifically covering facial contour descriptions such as round, oval, or square faces; descriptions of facial features such as large eyes, small nose, and thin lips; and other facial details, such as the presence of dimples or the location of moles. This detailed facial description text provides specific feature guidance for accurately generating the target subject's facial image, ensuring the generated image closely matches the target subject's actual facial features.

[0075] In this step, images can be filtered from a pre-built and organized image database, which contains a large amount of image data. Within the database, by setting search criteria related to the target object's blood relationship, such as father, mother, siblings, grandparents, maternal grandparents, etc., the corresponding target kinship images can be accurately located and extracted. Of course, other methods can also be used to obtain target kinship images; this is not a limitation in this case.

[0076] Meanwhile, when users have a clear understanding of the target object's facial features, they can describe these features in detail through the system's input interface, including facial contours, facial features, skin tone, and facial details. This information is then input into the system to obtain the target description text. Of course, other methods can also be used to obtain the target description text; this is not a limitation in this case.

[0077] Step 102: Extract target kinship features from the target kinship image.

[0078] The target kinship feature refers to the facial feature information of an individual who is related to the target object by blood, such as the facial genetic features exhibited by an individual who is related to the target object.

[0079] In this step, the human face can typically be divided into multiple structural units, such as eyes, nose, mouth, and cheeks, each carrying unique genetic information. Therefore, based on the principles of facial anatomy and visual structure division, the entire face can be divided into sub-images corresponding to multiple structural units. Examples include hair, forehead, left eye, right eye, left ear, right ear, nose, left cheek, right cheek, mouth, and chin. Then, for each structural unit, its corresponding feature extraction model and sub-image are obtained. The feature extraction unit processes the sub-image corresponding to that structural unit to obtain its features. For example, for the eye structural unit, its corresponding feature extraction unit can use edge detection algorithms to determine the eye contour, and then extract geometric features such as eye fissure length, interocular distance, and eyelid folds; simultaneously, it uses color analysis algorithms to obtain color and morphological features such as iris color and pupil size. For the nose structural unit, its corresponding feature extraction unit uses 3D reconstruction technology to obtain three-dimensional features such as nasal bridge height, nasal tip angle, and nasal wing width; and uses texture analysis algorithms to extract the texture details of the nose surface.

[0080] Alternatively, a first deep neural network is used to process the face image to obtain a feature map corresponding to the face image. Simultaneously, the face image is segmented to obtain a set of face region masks, which includes the face region mask corresponding to each structural unit. Then, the face region masks in the face region mask set are scaled to the same size as the aforementioned feature map, and the scaled masks are binarized to obtain binarized features, which in turn create a binarized feature set. For each binary feature corresponding to a structural unit in the binarized feature set, this binarized feature is multiplied by the aforementioned feature map to obtain a multiplication result. This multiplication result is then input into a second deep neural network to obtain the feature corresponding to that structural unit.

[0081] It should be noted that the neural network described above is trained on a large amount of data. It can be trained together with the models discussed later, or it can be trained using other methods. This is not a limitation here.

[0082] Based on the above method, after obtaining the features corresponding to each structural unit, these features are used as target kinship features.

[0083] In addition, if images of the target object when it was a child can be obtained in practice, the characteristics of the target object can be obtained from the image using the same method, and these characteristics can be used as target kinship characteristics. Then, the target genetic characteristics can be determined based on these target kinship characteristics.

[0084] Step 103: Detect whether target description features are extracted from the target description text.

[0085] Among them, the target description feature is the relevant feature describing the face of the target object. The target description feature is used to describe the facial feature information of the target object. For example, the target description feature is the facial features of the target object, or it can be the facial contour features of the target object.

[0086] In this step, the target description text is preprocessed. This includes removing noise from the text, such as irrelevant punctuation marks, special characters, and formatting information, making the text more organized and easier for subsequent processing. Then, pre-trained language models, such as BERT and GPT, are used. These models are trained on large-scale text and can learn rich semantic knowledge, thus better understanding the implicit semantic information in the processed text and extracting text features. Next, it is detected whether the text features are relevant to describing the target object's face. If so, the target description features are determined to be extracted from the target description text; otherwise, the target description features are determined not to be extracted.

[0087] Step 104: When the target description features are not extracted from the target description text, call the face generation model to process the target kinship features and obtain the face image of the target object.

[0088] Pre-trained face generation models are typically based on deep learning architectures, such as generative adversarial networks (GANs) or variational autoencoders (VAEs). During training, the model learns from a large amount of face image data, gradually mastering the complex mapping relationship between facial features and images.

[0089] In this step, when the pre-trained face generation model is invoked, the target kinship features are input to the model. Since the target kinship features carry facial feature information of people related to the target object by blood, when the model receives these features, it uses its internal neural network structure to perform deep processing to generate a face image that matches the input features.

[0090] It should be noted that, in order to improve the efficiency of model training and prediction, this application can also convert the target kinship features into discrete features before calling the face generation model to process the target kinship features and obtain the face image of the target object, and generate the face image of the target object based on the discrete features. The specific steps are as follows: for each structural unit, search for the discrete feature set corresponding to the structural unit in the face genetic feature library; use a similarity calculation algorithm to determine the similarity between each discrete feature in the discrete feature set and the target kinship features, and use it as the similarity corresponding to each discrete feature; determine the reference discrete features corresponding to the structural unit according to the discrete features whose similarity meets the preset conditions; after obtaining the reference discrete features corresponding to each structural unit, these reference discrete features can be fused to obtain fused features, and the face generation model can be called to process the fused features to obtain the face image of the target object.

[0091] The facial genetic feature database is a vast and meticulously organized collection of genetic feature information related to different facial structural units. These features are categorized and stored in a discrete form, covering a wide range of possible facial genetic feature types. For example, in the eye structural unit, it may include discrete features such as different shapes (round, almond-shaped, phoenix-shaped, etc.), eyelid types (single eyelid, double eyelid, inner double eyelid, etc.), and different iris colors. The preset conditions are based on genetic theory, practical research experience, and specific application needs. For example, preset conditions might be set to a similarity greater than a certain threshold (e.g., 0.7), or a high similarity ranking for certain discrete features.

[0092] In traditional deep learning-based face generation or style transfer techniques, facial features are typically mapped to a continuous high-dimensional vector space. When the model attempts to predict offspring features, it often employs weighted interpolation or regression of the parental feature vectors. However, this continuous space regression operation is prone to 'feature averaging' when dealing with highly nonlinear genetic features. For example, when the parents have 'high nose bridge' and 'flat nose bridge' respectively, the continuous vector model tends to generate a 'fuzzy intermediate state' feature between the two. This not only loses high-frequency texture details of the face, resulting in a distorted and overly smoothed facial image due to texture loss, but also fails to accurately reflect the discontinuous selection process of 'dominant / recessive' traits in genetics.

[0093] In contrast, the discretization encoding mechanism introduced in this application forces the predicted intermediate features to be mapped to a pre-constructed 'discrete feature set' consisting of clear and independent typical features. This is equivalent to introducing a 'quantization correction' process in the feature space: regardless of the intermediate results predicted by the model, the final determined target genetic feature must be a 'prototype feature' of a certain class that truly exists in the feature library and has clear texture. This mechanism effectively blocks the propagation of feature ambiguity, ensuring that each structural unit (such as eyes and nose) of the generated offspring face has highly realistic texture details and a clear morphological structure, thereby significantly improving the realism of face generation and the interpretability of genetic prediction.

[0094] In addition, the face generation system has a public face database. For each face image in the public face database, the following operations are performed to obtain the features corresponding to all structural units on each face. The specific steps include: processing the face image using a first deep neural network to obtain a feature map corresponding to the face image; simultaneously, dividing the face image to obtain a set of face region masks, which includes the face region mask corresponding to each structural unit. Then, scaling the face region masks in the face region mask set to the same size as the aforementioned feature map, and binarizing the scaled masks to obtain binary features, and then binarizing the feature set. Multiplying the binary features in the binarized feature set with the aforementioned feature map to obtain a multiplication result, and inputting this multiplication result into a second deep neural network to obtain its corresponding features. That is, obtaining the features corresponding to each structural unit on the face image. Then, for multiple face images, the features of the same structural unit are combined to obtain a feature set corresponding to each structural unit. For the feature set corresponding to each structural unit, a clustering method is used to divide the features in the feature set, obtaining multiple clustering results. Each clustering result essentially corresponds to a category. Therefore, the central feature in the clustering result is determined as the discrete feature corresponding to that category. The discrete features of all categories corresponding to this structural unit are combined to obtain its corresponding discrete feature set. Finally, the discrete feature set corresponding to each structural unit is used to form a facial genetic feature library.

[0095] Since the above method generates face images based on the reference discrete features corresponding to each structural unit, to make the generated face images more accurate, different weights can be assigned to the reference discrete features corresponding to key structural units and the reference discrete features corresponding to non-key structural units. Then, based on the reference discrete features and their corresponding weights, a fusion feature is obtained. Specifically, a first weight corresponding to the key structural unit and a second weight corresponding to the non-key structural unit are obtained. The first weight is then multiplied by the reference discrete feature corresponding to the key structural unit to obtain a new reference discrete feature corresponding to the key structural unit. The second weight is then multiplied by the reference discrete feature corresponding to the non-key structural unit to obtain a new reference discrete feature corresponding to the non-key structural unit. These new reference discrete features are then fused to obtain the fusion feature, and a face image is generated based on the fusion feature.

[0096] Non-critical structural units are structural units other than critical structural units.

[0097] The above fusion process can be a simple feature concatenation or a concatenation using a neural network; there is no limitation here.

[0098] Furthermore, the steps for determining key structural units are as follows: obtain the importance score of each structural unit for face recognition; and determine the structural units whose importance scores meet preset conditions as key structural units.

[0099] The importance score indicates the degree of importance of a structural unit to the face recognition process. The preset condition can be to identify structural units with an importance score greater than a preset score as key structural units, or to identify the largest preset number of structural units as key structural units, or other conditions; this is not limited here.

[0100] In this step, facial recognition experts assess the importance of each structural unit based on biological and psychological knowledge and practical experience, converting it into a score. Alternatively, a pre-trained machine learning model can be used to score each structural unit, obtaining a score for each unit. Other methods can also be used to determine the score for each structural unit. Then, structural units whose important scores meet preset conditions are identified and designated as key structural units.

[0101] In the above process, a pre-trained feature analysis model can be used to analyze the facial features of related people to obtain the target genetic features. Then, a pre-trained face generation model can be used to process the target genetic features to generate the target face image of the target object.

[0102] To train the aforementioned feature analysis model so that it can accurately extract genetic features from kinship images, the embodiments of this application employ the following training steps:

[0103] Step 1: Collect a large number of publicly available family photos or image data containing clear kinship relationships to construct a large-scale family kinship feature database, resulting in a sample dataset including family kinship relationships. Each sample in the sample dataset is a "family unit" containing at least one face image of the target object (children) and several face images of related relatives (such as father, mother, siblings, etc.).

[0104] Step 2: Using a pre-built feature extraction network (e.g., a deep convolutional neural network pre-trained on a large-scale face dataset or a publicly available face recognition model), process each image in each sample of the sample dataset. The feature extraction network acts as a feature encoder, mapping the input raw face image into a high-dimensional feature vector. This yields the baseline genetic feature vector corresponding to the target object in the sample, and the sample kinship feature vector corresponding to related objects.

[0105] Step 3: Construct a feature analysis model using the Transformer architecture. The model's input layer receives the sample kinship feature vectors corresponding to the aforementioned kinship objects. Utilizing the self-attention mechanism in the Transformer architecture, the model can capture the long-distance dependencies and association strengths between different sample kinship feature vectors, as well as between sample kinship feature vectors and potential genetic patterns.

[0106] Step 4: The feature analysis model also includes a weight predictor. The weight predictor connects to the Transformer module and, based on the contextual features output by the Transformer, predicts the contribution of each sample's kinship feature vector to the baseline genetic feature vector, outputting the corresponding feature weights. Subsequently, based on these feature weights, all sample kinship feature vectors are weighted and fused (e.g., weighted summation) to obtain the predicted genetic feature vector.

[0107] Step 5: During training, calculate the cosine similarity between the predicted genetic feature vector and the corresponding baseline genetic feature vector. Use the cosine similarity as the primary or sole metric of the loss function (e.g., loss = 1 - cosine similarity). Minimize this loss value using the backpropagation algorithm, continuously updating the parameters of the feature analysis model (including the Transformer module and weight predictor) until the model converges.

[0108] Through the above training method, the feature analysis model can learn how to accurately point to or approximate the true genetic characteristics of the target offspring in the vector space based on the combination of facial features of kinship objects, and finally obtain the trained feature analysis model.

[0109] When the feature analysis model is built based on statistical data methods, the specific steps for processing kinship facial features and determining the target genetic features of the target object using the feature analysis model are as follows: For each structural unit, in the genetic feature library, obtain the conditional probabilities of all first genetic feature combinations under the corresponding kinship facial features of that structural unit, obtain the conditional probability corresponding to each first genetic feature combination, and then determine the first genetic feature combination with the highest corresponding conditional probability, and determine it as the second genetic feature combination; based on the conditional probability corresponding to the second genetic feature combination and the kinship facial features, determine whether the feature in the second genetic feature combination is a strong genetic feature; when the feature in the second genetic feature combination is a strong genetic feature, determine the feature in the second genetic feature combination as the target genetic feature; when the feature in the second genetic feature combination is not a strong genetic feature, find the genetic feature combination with the highest correlation with the second genetic feature combination according to the genetic feature matrix, and determine the feature in the genetic feature combination as the target genetic feature.

[0110] The genetic feature matrix includes the correlation between combinations of genetic features, which are determined using genetic statistics methods. The genetic feature database stores all features, all combinations of genetic features, and the conditional probability of each combination of genetic features for each feature. A combination of genetic features is a combination of multiple genetic features.

[0111] The specific steps for determining whether a feature in the second genetic feature combination is a strong genetic feature combination based on the conditional probability corresponding to the second genetic feature combination and the kinship face features are as follows: detect whether the kinship face features are in the preset strong feature set and whether the conditional probability corresponding to the second genetic feature combination is greater than the dominant inheritance threshold. When both conditions are met, the feature in the second genetic feature combination is determined to be a strong genetic feature; otherwise, the feature in the second genetic feature combination is not a strong genetic feature.

[0112] For example, kinship facial features

[0113] in, Let represent the kinship feature corresponding to the i-th structural unit (hereinafter referred to as the i-th kinship feature), and n is the total number of structural units.

[0114] For the i-th kinship trait, its corresponding combination of second genetic traits ,in It is a set containing all combinations of features. For the j-th combination of genetic traits, Obtain the conditional probability corresponding to the j-th genetic feature combination under the i-th kinship feature.

[0115] Next, it checks whether the i-th kinship feature belongs to the strong feature set D, and then checks the second genetic feature combination. If the corresponding conditional probability is greater than the dominant inheritance threshold, then the second genetic trait is combined. The characteristics identified in the text are strong heritable traits. Otherwise, the genetic trait combination with the strongest association with the second genetic trait combination will be selected, and the genetic trait in that combination will be identified as the target genetic trait. Specifically, the formula is as follows:

[0116]

[0117] in, The combination of genetic characteristics in the genetic transition matrix Combination of genetic traits The strength of the association between two genetic traits is the probability / correlation of a combination of genetic traits transforming into, retaining, or co-occurring with another combination of genetic traits during the genetic process (such as reproduction, iteration, natural selection / algorithmic evolution). It can be obtained through statistical analysis of a large amount of genetic data, or through other methods, which will not be elaborated here.

[0118] When the feature analysis model is a machine learning model, the specific steps for processing kinship facial features and determining the target genetic characteristics of the target object are as follows: During the analysis of kinship facial features, the pre-trained feature analysis model performs a comprehensive analysis of the kinship facial features. Through this comprehensive and in-depth analysis, the model ultimately determines the target genetic characteristics that represent the genetic traits of the target object.

[0119] The pre-trained feature analysis model is built upon a large amount of facial genetic data and machine learning algorithms. This model aims to extract deep-seated genetic information from the kinship facial features corresponding to each structural unit, thereby determining the target genetic features.

[0120] To extract target genetic features more accurately, this application can set different feature analysis models for different combinations of kinship relationships. This allows for the initial determination of the kinship relationship between each individual and the target individual, followed by the determination of the feature analysis model to be used based on the combination of these kinship relationships. For example, one feature analysis model corresponds to the father and mother, and another to the father and his siblings.

[0121] For example, when the father and mother each correspond to a feature analysis model, the method for obtaining the pre-trained feature analysis model is as follows: Assume the father's face region encoding feature is F_r, the mother's face region encoding feature is M_r, and the offspring's face region encoding feature to be predicted is S_r. Select several parent-child pairing data to construct a training dataset with the parents' encoding features as input and the offspring's encoding features as labels. Concatenate the father's face region encoding feature F_r and the mother's face region encoding feature M_r to obtain a fused feature P_r. Input the fused feature into the feature analysis model, and the output is the probability vector V_r for each region of the offspring. Perform One-Hot encoding on the offspring's face region encoding feature S_r to obtain labels. Train the feature analysis model by calculating the cross-entropy loss between the probability vector V_r and the One-Hot encoded labels. Finally, the pre-trained feature analysis model is obtained.

[0122] In the above process, when there are multiple target kinship features, the kinship relationship between the target kinship feature and the target object can be obtained, the weight corresponding to each kinship object can be determined, and the target kinship feature and the corresponding weight can be input into the feature analysis model so that the genetic feature analysis model can analyze and process the input data to obtain the target genetic feature.

[0123] The aforementioned feature analysis model is built upon a large amount of facial genetic data and machine learning algorithms. This model aims to extract deep-seated genetic information from the target kinship features corresponding to each structural unit, thereby determining the target genetic characteristics.

[0124] When the kinship relationship between the kinship object and the target object is the first kinship relationship, the weight corresponding to the kinship object is set as the first weight; when the kinship relationship between the kinship object and the target object is the second kinship relationship, the weight corresponding to the kinship object is set as the second weight; when the kinship relationship between the kinship object and the target object is the third kinship relationship, the remaining total weight is determined according to the first and second weights; the ratio of the remaining total weight to the number is determined as the third weight, and the weight corresponding to the kinship object is set as the third weight.

[0125] The first kinship relationship is the relationship between the mother and the target. The second kinship relationship is the relationship between the father and the target. The third kinship relationship is the relationship between the siblings and the target.

[0126] In this step, for each related object, when the kinship between the related object and the target object is the first kinship, the weight corresponding to the related object is set as the first weight. When the kinship between the related object and the target object is the second kinship, the weight corresponding to the related object is set as the second weight. When the kinship between the related object and the target object is the third kinship, the remaining total weight is determined based on the first and second weights. Subsequently, the ratio of the remaining total weight to the corresponding number is calculated to determine the third weight, and the weight corresponding to the related object is set as the third weight. The specific steps for determining the remaining total weight based on the first and second weights are as follows: since the sum of all weights is 1, 1 is subtracted from the first weight and the second weight, and the difference is determined as the remaining total weight.

[0127] For example, if the weight corresponding to the first kinship is 0.3, the weight corresponding to the second kinship is 0.3, the weight corresponding to the third kinship is 0.35, and the number of kinship objects corresponding to the third kinship is 5, then the ratio of the third weight of 0.35 to 5 is 0.7.

[0128] Step 105: When extracting target description features from the target description text, detect whether the target description features include genetic description features and facial description features, obtain the detection results, call the face generation model, process the target kinship features and the features in the detection results, and obtain the face image of the target object.

[0129] Among them, genetic descriptive features are heritable facial characteristics within the target individual's family, such as common family features like a high nose bridge, deep-set eyes, and specific ear shapes. It also includes descriptions of physiological characteristics with heritable predispositions, such as skin color, hair color, and eye color. Facial descriptive features are unique facial details of the target individual, such as the location and shape of moles, birthmarks, and dimples, as well as facial expression habits, such as the upward curve of the corners of the mouth when smiling and the lines formed when frowning.

[0130] In this step, when extracting target description features from the target description text, it is detected whether the target description features include genetic description features and facial description features, and a detection result is obtained. If the detection result indicates that the target description features include both genetic and facial description features, a face generation model is invoked to process the target kinship features, genetic description features, and facial description features to obtain the face image of the target object. If the detection result indicates that the target description features only include genetic description features, a face generation model is invoked to process the target kinship features and genetic description features to obtain the face image of the target object. If the detection result indicates that the target description features only include facial description features, a face generation model is invoked to process the target kinship features and facial description features to obtain the face image of the target object.

[0131] In this embodiment, a target kinship image and a target description text are obtained. The target kinship image is an image of an object related to the target by blood, and the target description text includes facial description information of the target object. Target kinship features are extracted from the target kinship image. It is detected whether target description features are extracted from the target description text. When no target description features are extracted from the target description text, a face generation model is invoked to process the target kinship features to obtain a face image of the target object. When target description features are extracted from the target description text, it is detected whether the target description features include genetic description features and facial description features. A detection result is obtained, and a face generation model is invoked to process the target kinship features and the features in the detection result to obtain a face image of the target object. This application can generate a face image of the target object based on an image of the target object's relatives, combined with a textual description of the target object. The generated face image can highly realistically reflect the current facial features of a missing or maliciously transferred young object, providing crucial and powerful evidence for handling such cases.

[0132] In this embodiment, a pre-trained face generation model can fuse target kinship features with features from the detection results to obtain fused features. Finally, based on these fused features, a face image of the target object is generated. Therefore, the face generation model can be divided into a fusion module and a first generation module. The fusion module fuses target kinship features with features from the detection results to obtain fused features, which are then input into the first generation module. After processing by this module, a face image of the target object is generated. Therefore, this embodiment provides a face image generation method, as follows: Figure 2 As shown, the specific steps include:

[0133] Step 201: Use the fusion module to fuse the target kinship features and the features in the detection results to obtain fused features.

[0134] Among them, the fusion feature is obtained by integrating multiple features and is used to generate a face image based on the fusion feature.

[0135] In this step, the target kinship features and the features from the detection results are input into the fusion module. The fusion module can concatenate the target kinship features and the features from the detection results to obtain the fused features. It can also obtain the weights corresponding to the target kinship features and the target descriptive features, multiply the target kinship features and the features from the detection results by their respective weights, and concatenate the multiplication results to obtain the fused features. Alternatively, the fusion module can also be a convolutional neural network (CNN) or a recurrent neural network (RNN), which can automatically learn how to fuse features through the training process.

[0136] Step 202: Input the fused features into the first generation module so that the first generation module processes the fused features to obtain the face image of the target object.

[0137] In this step, the first generation module is a meticulously designed and trained machine learning module with powerful feature processing and image generation capabilities. When the fused features are input into the first generation module, it performs in-depth processing based on its internally defined complex algorithms and deep neural network architecture. During processing, the first generation module parses, transforms, and reconstructs the fused features, gradually converting abstract feature information into visualized image elements. Finally, by continuously adjusting the image's pixel values, color distribution, and texture details, the generated image gradually approximates the target object's face image, thus obtaining the target object's face image.

[0138] In this embodiment, the target description text may include genetic description text and facial description text. The genetic description text includes information about hereditary facial features within the target's family, such as a high nose bridge, deep-set eyes, and specific ear shapes commonly found in the family. It also includes descriptions of hereditary physiological characteristics such as skin color, hair color, and eye color. The facial description text includes unique facial details of the target, such as the location and shape of moles, birthmarks, and dimples, as well as facial expression habits, such as the upward curve of the corners of the mouth when smiling and the lines formed when frowning. Therefore, the target description text can be identified to detect whether it includes both genetic and facial description text. If both are included, the target description text is identified, separating the genetic and facial description texts. Then, the genetic and facial description texts are identified separately to obtain the corresponding genetic and facial description features. Therefore, this application provides a target description text recognition method, the specific steps of which include: recognizing the target description text to obtain genetic description text and face description text; recognizing the genetic description text to obtain genetic description features; and recognizing the face description text to obtain face description features.

[0139] In this step, natural language processing techniques and algorithms are used to accurately segment different types of descriptive text by conducting in-depth analysis of the target descriptive text at the lexical, grammatical, and semantic levels, resulting in genetic descriptive text and facial descriptive text. Then, for the genetic descriptive text, a first text recognition model trained based on genetics knowledge is used to deeply mine and understand the genetic information-related words and sentences. Genetically significant feature information is extracted from this text content, thus obtaining genetic descriptive features. Simultaneously, for the facial descriptive text, a second text recognition model combining computer vision and natural language processing is used to transform the descriptions of facial features in the text into quantifiable and analyzable information, thereby obtaining facial descriptive features.

[0140] In this embodiment, when the detection result indicates that the target description text includes both genetic description text and facial description text, genetic description features and facial description features can be extracted from the target description text; that is, the target description features include both genetic description features and facial description features. In this case, the face generation model includes a second generation module and an image modification model. The second generation module processes the target kinship features and genetic description features to obtain a reference face image. Then, the image modification module further processes the facial description features and the reference face image to finally obtain the face image of the target object. Therefore, this embodiment provides a method for determining a reference face image, as follows: Figure 3 As shown, the specific steps include:

[0141] Step 301: Use the second generation module to process the target kinship features and genetic description features to obtain a reference face image.

[0142] In this step, the second generation module is built on a deep learning architecture such as a convolutional neural network (CNN). When the target's kinship features and genetic description features are input into this module, the module initiates a series of complex neural network-based computations, ultimately outputting a reference face image.

[0143] In this step, when inputting the target kinship features and genetic description features into the second generation module, the target kinship features and genetic description features can also be fused to obtain fused features, which are then input into the second generation model. Since the target kinship features include features corresponding to each structural unit, the genetic description features can also be analyzed to determine their associated structural units. The features corresponding to the associated structural units are then fused with the genetic description features to obtain the fused features corresponding to the associated structural units. Finally, the features corresponding to all structural units are fused to obtain the fused features. The above fusion process can be a simple concatenation of the target kinship features and genetic description features, or other fusion processes; it is not limited here.

[0144] Specifically, in the process of fusing the features corresponding to the associated structural units and the genetic descriptive features, the weights corresponding to the features and the genetic descriptive features can be obtained. The weights corresponding to the features are multiplied by the features to obtain a multiplication result, and the weights corresponding to the genetic descriptive features are multiplied by the genetic descriptive features to obtain another multiplication result. The two multiplication results are then superimposed to obtain the fused feature.

[0145] Step 302: Use the image modification module to process the facial description features and the reference facial image to obtain the facial image of the target object.

[0146] In this step, the image modification module uses advanced image processing algorithms to comprehensively process both facial descriptive features and a reference facial image. Guided by facial descriptive features, this module deeply analyzes the differences between various parts of the reference facial image and the facial descriptive features to obtain the target object's facial image. For example, if the facial descriptive features indicate that the target object has a wide forehead, while the forehead of the reference facial image is relatively narrow, the image modification module will use a specific algorithm to stretch or deform the forehead portion of the reference facial image to better match the forehead features of the target object.

[0147] In this embodiment, when the detection result indicates that the target description text only includes genetic description text, the second generation module can be directly used to process the target kinship features and genetic description features to obtain the target object's face image. Alternatively, the second generation module can be directly used to process the target kinship features to obtain the target object's face image, and finally, the image modification module can be used to process the genetic description features and the reference face image to obtain the target object's face image. When the target description text only includes face description features, the method is similar to the above, and will not be elaborated here.

[0148] In this embodiment, the target text can also be processed directly using the second generation module to obtain a reference face image without differentiation. Then, an image modification module is used to process the target description features and the reference face image to obtain the face image of the target object.

[0149] In this embodiment, a first weight of the target kinship feature and a second weight of the genetic descriptive feature are obtained. Based on these two weights, the target kinship feature and the genetic descriptive feature are fused to obtain a first fused feature, which is then processed to obtain a reference face image. The genetic descriptive feature may contain broader genetic information, while the target kinship feature focuses on specific kinship characteristics. By adjusting the second weight, a suitable balance can be found between the two, ensuring that the universality of genetic information and the specificity of kinship are both reflected in the first fused feature. Therefore, this embodiment provides a method for generating a reference face image. The specific steps of this method include: obtaining the first weight corresponding to the target kinship feature and the second weight corresponding to the genetic descriptive feature; fusing the target kinship feature and the genetic descriptive feature according to the first and second weights to obtain the first fused feature; and processing the first fused feature to obtain a reference face image.

[0150] In this step, firstly, the first weight corresponding to the target kinship feature and the second weight corresponding to the genetic descriptive feature are obtained. These weights typically rely on specific algorithms or are based on relevant domain knowledge and experience. For example, in some kinship studies based on big data analysis, the algorithm calculates the corresponding weights of the target kinship feature and the genetic descriptive feature based on their importance in practical application scenarios through learning and analysis of a large amount of sample data. These weights reflect the relative importance of different features in the overall information. Next, based on the obtained first and second weights, the target kinship feature and the genetic descriptive feature are fused. This fusion process can be a simple superposition or based on a specific mathematical model or fusion algorithm. For example, a weighted summation method might be used, multiplying the target kinship feature by the first weight and the genetic descriptive feature by the second weight, and then adding them together to obtain the first fused feature. The first fused feature integrates the key information of the target kinship feature and the genetic descriptive feature, while preserving their respective weight ratios in the overall picture. Finally, the first fused feature is processed to obtain a reference face image. The processing may involve complex image processing algorithms, machine learning models, or deep learning architectures. For example, using a Generative Adversarial Network (GAN), the first fused features are taken as input, and the generated image is gradually optimized through adversarial training between the generator and the discriminator so that it visually presents a face shape that conforms to the description of these features, and finally a reference face image is generated.

[0151] In this embodiment, the weights corresponding to the face description features and the reference face image are first obtained. Then, the face description features and the reference face image are fused according to these two weights to generate a second fused feature. Finally, the face image of the target object is obtained by processing the second fused feature. When facing different application scenarios and needs, the generated face image can better meet the requirements of specific scenarios by flexibly adjusting the third and fourth weights. Therefore, this embodiment provides a face image determination method, which specifically includes: obtaining the third weight corresponding to the face description features and the fourth weight corresponding to the reference face image; fusing the face description features and the reference face image according to the third and fourth weights to obtain a second fused feature; and processing the second fused feature to obtain the face image of the target object.

[0152] The third weight is used to highlight the unique personalized information of the target object, such as special facial markings and expression habits. When the third weight is high, these personalized features dominate the fused features, making the generated facial image more closely match the individual characteristics of the target object. The fourth weight is used to strengthen the overall facial style constructed based on the target's kinship features and genetic description features. A higher fourth weight ensures that common family features or basic features generated based on genetic information are highlighted in the final image. The third and fourth weights can be set by technicians based on experience, or determined using other methods; this is not limited here.

[0153] This step is similar to the embodiment for generating a reference face image, and is not limited thereto.

[0154] In this embodiment, the presence of associated genetic features of the target kinship feature in the genetic feature database is detected. When associated genetic features of the target kinship feature exist in the genetic feature database, a face generation model is invoked to process the target kinship feature and the associated genetic features to obtain the face image of the target object. Therefore, this embodiment provides a face image determination method, as follows: Figure 4 As shown, the specific steps include:

[0155] Step 401: Detect whether there are related genetic features of the target kinship in the genetic feature database.

[0156] Among them, the related genetic features of the target kinship feature refer to the set of features that are closely related to the target kinship feature, determined by genetic factors, and can further supplement or refine the facial feature information of the target object. For example, the high brow bone feature that grandparents have may not be expressed in the parents' generation, but may appear in the grandchildren. This high brow bone feature that is inherited across generations belongs to the related genetic features of the target kinship feature (kinship feature between parents and children).

[0157] In this step, the system will conduct a comprehensive search of the genetic feature database targeting the target kinship feature. The genetic feature database stores a massive amount of genetic feature data and the relationships between them. During the detection process, the system will use the target kinship feature as a clue to perform a detailed search and matching within the genetic feature database to determine if any related genetic features exist.

[0158] Step 402: When the genetic feature library contains related genetic features of the target kinship feature, call the face generation model to process the target kinship feature and related genetic features to obtain the face image of the target object.

[0159] Among them, face generation models are usually complex models built on deep learning algorithms. They are trained on a large amount of face data and have powerful feature processing and image generation capabilities.

[0160] In this step, if the related genetic features of the target kinship feature are successfully detected in the genetic feature library, the subsequent key steps will be triggered. At this time, the system inputs the target kinship feature and the related genetic features into the face generation model, so that the face generation model processes the target kinship feature, the related genetic features, and the target descriptive features to obtain the face image of the target object.

[0161] It should be noted that when there are no related genetic features of the target kinship feature in the genetic feature library, the face generation model is called to process the target kinship feature and the target descriptive feature to obtain the face image of the target object.

[0162] In addition, the system detects whether there are related genetic features of the target kinship in the genetic feature database. When such features exist, the face generation model is invoked to process the target kinship, related genetic features, and features from the detection results to obtain the face image of the target object. This embodiment is similar to the above embodiments and will not be described in detail here.

[0163] In this embodiment of the application, after obtaining the face image of the target object, the target object is searched in the public face database based on the face image. The specific steps are as follows: using a similarity calculation algorithm, the similarity between each face in the public face database and the newly generated adult face image is calculated, and a preset number of faces with the highest similarity are taken as faces similar to the newly generated adult face image, and the object corresponding to the face image is determined as the target object.

[0164] Furthermore, when multiple target objects are identified, the household registration information of each target object is obtained; for each target object, the corresponding household registration score is determined based on the corresponding household registration information; and based on the corresponding household registration score of each target object, the target objects that meet the preset conditions are determined among all target objects.

[0165] The higher the household registration score, the better the public security in the corresponding area, and the lower the probability of related adverse events. Conversely, the lower the household registration score, the worse the public security in the corresponding area, and the higher the probability of related adverse events. For example, if the target's household registration information is from location A, the corresponding household registration score can be set to 9 because location A has good public security. If the target's household registration information is from location A, the corresponding household registration score can be set to 4 because location A has experienced many related adverse events. The aforementioned related adverse events refer to human trafficking. The preset conditions can be objects with household registration scores lower than a preset score or a preset number of objects with the lowest household registration scores. Of course, the preset conditions can also be other conditions, which are not limited here.

[0166] In this step, after identifying the target objects, for each target object, the household registration information of the target object is queried. Based on the pre-set correspondence between household registration information and household registration scores, the corresponding household registration score is determined and set as the household registration score for that target object. Finally, based on the household registration score corresponding to each target object, the target objects that meet the preset conditions are identified from all target objects.

[0167] In this embodiment, the target object can also be verified to determine if it is more likely to be the object being searched. Specific steps include: obtaining a preset verification strategy; and using the preset verification strategy to verify the target object.

[0168] In this step, a preset verification strategy can generate a verification notification for the system and send it to the relevant department in the target's household registration office, enabling the department to verify the target's information. For example, the verification notification could involve collecting DNA samples from the target and their current family members to determine if a blood relationship exists between them.

[0169] Based on the above embodiments, after obtaining a facial image, since the facial image excludes objective attribute differences such as gender and age, searching for targets based on this facial image can more accurately and efficiently locate individuals matching the facial features, reducing interference from non-core facial feature differences such as gender and age, improving the accuracy and targeting of the search, and finding the target object faster. Specific solutions include:

[0170] Case 1: During the initial facial comparison of suspected individual A (out of 800 people), over 100 individuals with "similar appearances" were manually selected. However, due to A's young age at the time of disappearance and significant changes in appearance as an adult (age interference), even though A was on the candidate list, he was deemed "low probability" and shelved. Subsequently, the team used cross-age kinship comparison technology to eliminate age interference and re-identified suspected individual A. Ultimately, after two DNA verifications, it was confirmed that the man was indeed suspected individual A, who had been missing for 22 years.

[0171] Case 2: Suspect B was transferred to Region 1 at the age of 3. After reaching adulthood, he returned to Region 2 with his adoptive parents, while his biological parents remained in Region 2. Due to "differences in appearance between childhood and adulthood (age interference)," the two groups repeatedly missed each other without being identified. In 2023, the authorities, using facial recognition technology that excludes age- and gender-irrelevant features, successfully matched the adult face of suspect B with the face of the child who went missing, ultimately leading to their reunion on September 26.

[0172] Case 3: From 1993 to 1996, suspect XX and his accomplice XXX traveled to various locations, renting houses and familiarizing themselves with the local environment to target suitable young children for malicious relocation. They repeatedly used these methods to maliciously relocate 11 young children, including suspected individuals C and D, to region 3. Furthermore, of the 13 people suspected of being maliciously relocated by suspect XX, 10 have been reunited with their families of origin, 2 have been located by their suspected relatives, and 1 remains missing.

[0173] Case 4: Suspect F was maliciously moved when he was only 100 days old, and his parents searched for him for 25 years without success. After the authorities apprehended the suspect in 2022, they only learned that suspect F had been maliciously moved to a location near region 4. However, due to the large time span and the fact that suspect F's appearance had completely changed beyond infancy (age interference), the search reached a stalemate. In November 2023, the authorities used cross-age kinship comparison technology (using the core facial features of suspect F's parents as anchor points to eliminate interference from suspect F's age and gender) to quickly identify the suspect. After DNA confirmation, the parents were reunited on December 1st.

[0174] Beyond specific case studies, various data and charts also visually demonstrate the powerful effectiveness of the technology. Table 1 presents the number of matches using only early childhood characteristics and combined with same-kinship characteristics for different ranges of "topk" candidate numbers, clearly showing that same-kinship comparison can effectively improve the match rate, specifically:

[0175] Table 1

[0176]

[0177] Figure 5 The bar chart shows a significant increase in the number of cases involving the trafficking of minors solved each year, which indirectly confirms the significant assistance that related technologies have in solving these cases. Figure 6 The bar chart shows the number of times the ranking of the 2 million database improved under different original database ranking ranges, reflecting the effect of technology on improving the ranking of search objects.

[0178] Meanwhile, looking at historical cases: 1. Mr. Chen: Using traditional methods, it took 2 years without results; using the same kinship method, the algorithm results were analyzed in less than 1 hour; 2. Xiao C: Less than 1 hour; 3. Xiao Xie: Less than 1 hour; and some even got results in 0.5 hours.

[0179] It should be understood that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order constraint on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the diagram may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0180] Please see Figure 7 One embodiment of this application provides a face image generation device 700 based on genetic laws, comprising:

[0181] The acquisition unit 701 is used to acquire a target kinship image and the target description text, wherein the target kinship image is an image of an object that is related to the target object by blood, and the target description text includes facial description information of the target object;

[0182] Extraction unit 702 is used to extract target kinship features from the target kinship image, wherein the target kinship features are facial feature information of objects that are related to the target object by blood;

[0183] Detection unit 703 is used to detect whether target description features are extracted from the target description text, wherein the target description features are used to describe the facial feature information of the target object;

[0184] The first calling unit 704 is used to call a face generation model to process the target kinship features and obtain the face image of the target object when the target description features are not extracted from the target description text.

[0185] The second calling unit 705 is used to detect whether the target description features include genetic description features and face description features when extracting target description features from the target description text, obtain the detection result, call the face generation model, process the target kinship features and the features in the detection result, and obtain the face image of the target object.

[0186] Optionally, the face generation model includes a fusion module and a first generation module, and a second calling unit 705, used for:

[0187] The fusion module is used to fuse the target kinship features and the features in the detection results to obtain fused features;

[0188] The fusion features are input into the first generation module so that the first generation module processes the fusion features to obtain the face image of the target object.

[0189] Optionally, the detection result includes genetic descriptive features and facial descriptive features, the facial generation model includes a second generation module and an image modification model, and the second calling unit 705 is used for:

[0190] Using the second generation module, the target kinship features and the genetic description features are processed to obtain a reference face image;

[0191] The image modification module is used to process the facial description features and the reference facial image to obtain the facial image of the target object.

[0192] Optionally, the device further includes an identification unit 706, the identification unit 706 being used for:

[0193] The target description text is identified to obtain genetic description text and face description text;

[0194] The genetic description text is identified to obtain the genetic description features, and the facial description text is identified to obtain the facial description features.

[0195] Optionally, the second calling unit 705 is used for:

[0196] Obtain the first weight corresponding to the target kinship feature and the second weight corresponding to the genetic description feature;

[0197] Based on the first weight and the second weight, the target kinship feature and the genetic description feature are fused to obtain a first fused feature;

[0198] The first fused feature is processed to obtain a reference face image.

[0199] Optionally, the first calling unit 704 is used for:

[0200] Detect whether the target kinship characteristic is associated with a genetic feature in the genetic feature database;

[0201] When the genetic feature library contains related genetic features of the target kinship feature, the face generation model is invoked to process the target kinship feature and the related genetic features to obtain the face image of the target object.

[0202] Optionally, the second calling unit 705 is used for:

[0203] Obtain the third weight corresponding to the face description feature and the fourth weight corresponding to the reference face image;

[0204] Based on the third weight and the fourth weight, the target kinship feature and the genetic description feature are fused to obtain the second fused feature;

[0205] The second fusion feature is processed to obtain the face image of the target object.

[0206] Specific limitations regarding the aforementioned device for generating face images based on genetic principles can be found in the limitations of the method for generating face images based on genetic principles described above, and will not be repeated here. Each unit in the aforementioned device for generating face images based on genetic principles can be implemented entirely or partially through software, hardware, or a combination thereof. These units can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0207] In one embodiment, a computer device is provided, the internal structure of which can be as follows: Figure 8 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and the database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores data. The network interface communicates with external terminals via a network connection. The computer program, executed by the processor, can implement the above-described method for generating human face images based on genetic principles. It includes: memory and a processor; the memory stores the computer program; and the processor executes the computer program to implement any step in the above-described method for generating human face images based on genetic principles.

[0208] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, can perform any of the steps in the above method for generating face images based on genetic laws.

[0209] Those skilled in the art will understand that embodiments of this application can be provided as a method, system, or computer program product for generating facial images based on genetic principles. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0210] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0211] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0212] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0213] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0214] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for generating human face images based on genetic principles, characterized in that, include: Obtain the target kinship image and the target description text, wherein the target kinship image is an image of an object that is related to the target object by blood; Extract target kinship features from the target kinship image, wherein the target kinship features are facial feature information of objects that are related to the target object by blood; Detect whether target description features are extracted from the target description text, wherein the target description features are used to describe the facial feature information of the target object; When the target description features are not extracted from the target description text, the face generation model is invoked to process the target kinship features to obtain the face image of the target object; When extracting target description features from the target description text, it is detected whether the target description features include genetic description features and facial description features. A detection result is obtained, and a face generation model is invoked to process the target kinship features and the features in the detection result to obtain the face image of the target object. The face generation model includes a fusion module and a first generation module. Invoking the face generation model to process the target kinship features and the features in the detection result to obtain the face image of the target object includes: The fusion module is used to fuse the target kinship features and the features in the detection results to obtain fused features; the fused features are then input into the first generation module so that the first generation module processes the fused features to obtain the face image of the target object.

2. The method according to claim 1, characterized in that, The detection result indicates that the target descriptive features include genetic descriptive features and facial descriptive features. The face generation model includes a second generation module and an image modification module. Calling the face generation model to process the target kinship features and the features in the detection result to obtain the face image of the target object includes: Using the second generation module, the target kinship features and the genetic description features are processed to obtain a reference face image; The image modification module is used to process the facial description features and the reference facial image to obtain the facial image of the target object.

3. The method according to claim 2, characterized in that, The method further includes: The target description text is identified to obtain genetic description text and face description text; The genetic description text is identified to obtain the genetic description features, and the facial description text is identified to obtain the facial description features.

4. The method according to claim 2, characterized in that, The process of processing the target kinship features and the genetic description features to obtain a reference face image includes: Obtain the first weight corresponding to the target kinship feature and the second weight corresponding to the genetic description feature; Based on the first weight and the second weight, the target kinship feature and the genetic description feature are fused to obtain a first fused feature; The first fused feature is processed to obtain a reference face image.

5. The method according to claim 1, characterized in that, The process of calling the face generation model to process the target kinship features and obtain the face image of the target object includes: Detect whether the target kinship characteristic is associated with a genetic feature in the genetic feature database; When the genetic feature library contains related genetic features of the target kinship feature, the face generation model is invoked to process the target kinship feature and the related genetic features to obtain the face image of the target object.

6. The method according to claim 2, characterized in that, The process of processing the facial description features and the reference facial image to obtain the facial image of the target object includes: Obtain the third weight corresponding to the face description feature and the fourth weight corresponding to the reference face image; Based on the third weight and the fourth weight, the face description features and the reference face image are fused to obtain the second fused feature; The second fusion feature is processed to obtain the face image of the target object.

7. A device for generating human face images based on genetic principles, characterized in that, The device includes: The acquisition unit is used to acquire a target kinship image and a target description text, wherein the target kinship image is an image of an object that is related to the target object by blood, and the target description text includes facial description information of the target object; The extraction unit is used to extract target kinship features from the target kinship image, wherein the target kinship features are facial feature information of objects that are related to the target object by blood; A detection unit is used to detect whether target description features are extracted from the target description text, wherein the target description features are used to describe the facial feature information of the target object; The first calling unit is used to call a face generation model to process the target kinship features and obtain the face image of the target object when the target description features are not extracted from the target description text. The second calling unit is used to, when extracting target description features from the target description text, detect whether the target description features include genetic description features and facial description features, obtain a detection result, and call a face generation model to process the target kinship features and the features in the detection result to obtain a face image of the target object; wherein, the face generation model includes a fusion module and a first generation module, and the step of calling the face generation model to process the target kinship features and the features in the detection result to obtain a face image of the target object includes: The fusion module is used to fuse the target kinship features and the features in the detection results to obtain fused features; the fused features are then input into the first generation module so that the first generation module processes the fused features to obtain the face image of the target object.

8. A computer device, comprising: A memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images