AI recognition method and device for face image splitting and recombining and application

Abstract

The invention provides an AI recognition method and device for face image splitting and recombining and application, and relates to the technical field of image information processing. The processing method comprises the following steps: acquiring picture information containing a face image of an object to be recognized; extracting facial information of a to-be-recognized object from at least one frame of picture information, and judging whether at least one facial feature which does not belong to the to-be-recognized object exists in the facial information or not through AI recognition; and if yes, determining the identity information of a person to which the facial feature actually belongs by taking at least one facial feature which does not belong to the object to be recognized as a reference. According to the invention, the face information splitting and recombining behaviors of the to-be-recognized object can be recognized, the identity information of the personnel corresponding to the face features which do not belong to the to-be-recognized object is determined, and how the image information of the to-be-recognized object is obtained by splitting and recombining the face features of others is learned; and therefore, fraudulent behaviors executed through face image splitting and recombining behaviors in the network can be identified.

Description

Technical Field

[0001] The present invention relates to the field of image information processing technology, and in particular to an AI recognition method for splitting and recombining facial images. Background Art

[0002] With the rapid development and widespread application of facial recognition technology, there has been an increasing number of attempts to deceive people's facial information or to gain publicity online. These attempts may exploit the disassembly and reassembly properties of facial information, changing facial features or imitating others' facial features to mislead or deceive facial recognition systems.

[0003] For example, face synthesis technology can generate a fake facial image by splitting and reassembling the facial features of different people. This technology can be used to forge identities to deceive facial recognition systems or create false personal images on social media.

[0004] For example, video face-changing technology can extract a person's facial features from a video, and then split and recombine them with the facial features of another person to achieve the exchange of facial features. This technology can be used to create fake videos to make people have wrong perceptions.

[0005] The existence of these technologies makes it easier to deceive or create hype around people's facial information. Therefore, in this situation, it is very important to strengthen the ability to resist deception attacks and accurately identify whether facial features in images containing human faces are split or reassembled.

[0006] To this end, the present application provides an AI recognition method, device and application for splitting and recombining facial images to identify the facial information of the object to be identified that has been split and recombined, which is a technical problem that urgently needs to be solved. Summary of the Invention

[0007] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide an AI recognition method, device and application for face image splitting and recombining. The present invention can identify the behavior of splitting and recombining the facial information of the object to be identified, determine the identity information of the person corresponding to the facial features that do not belong to the object to be identified, and learn how the image information of the object to be identified is obtained by splitting and recombining the facial features of others, thereby identifying fraudulent behavior performed on the network through the behavior of splitting and recombining face images.

[0008] In order to solve the existing technical problems, the present invention provides the following technical solutions: An AI recognition method for splitting and recombining facial images, comprising the steps of: Acquire image information containing a face image of a subject to be identified; the image information includes video image information and image image information; the video image information is composed of multiple frames of continuous image image information; Extracting facial information of the subject to be identified from at least one frame of the image information, and determining through AI recognition whether the facial information contains at least one facial feature that does not belong to the subject to be identified; wherein the facial feature that does not belong to the subject to be identified refers to at least one facial feature that does not belong to the subject to be identified and, after being split and reassembled, becomes at least one facial feature presented in the facial image of the subject to be identified in the image information; When the judgment is yes, the identity information of the person to whom the facial feature actually belongs is determined based on at least one facial feature that does not belong to the aforementioned object to be identified.

[0009] Furthermore, the image information of the face image of the object to be identified refers to detailed visual data containing individual facial features extracted from a face photo or video; The image information includes facial geometry, appearance, expression changes, lighting conditions, posture information, environmental information, time information, resolution and quality; The facial geometric features refer to the facial geometric structure presented by the object to be identified; Appearance features refer to facial characteristics that can be directly observed visually, including the skin color, skin texture, facial lines, facial features, facial proportions, facial symmetry, wrinkles, creases, scars, moles, spots, facial expressions, makeup and jewelry of the face of the subject to be identified; The expression change refers to the dynamic change of the face of the object to be identified; The lighting conditions refer to the direction, intensity and color temperature of light that affect the quality of the facial image and the visibility of features of the object to be identified; The posture information refers to the angle of the face of the object to be identified, and the posture information includes looking down, looking up and side face; The environmental information refers to the objects and light conditions in the background of the object to be identified; The time information includes timestamp information corresponding to at least one frame of image information involving the facial image of the object to be identified in a video or dynamic monitoring environment; the timestamp information indicates the specific capture time of the facial image of the object to be identified; The resolution and quality refer to the clarity and pixel size of the facial image of the object to be identified, as well as any compression or damage that may affect the accuracy of recognition.

[0010] Furthermore, the facial features that do not belong to the aforementioned object to be identified also include facial features corresponding to the object to be identified at different ages that are split and reassembled to become at least one facial feature presented in the face image of the object to be identified in the aforementioned picture information.

[0011] Furthermore, when extracting the facial information of the object to be identified, the extraction operation is performed only on the facial information containing the facial features of the object to be identified in the aforementioned image information.

[0012] Furthermore, the determination includes the steps of: Using a facial feature extraction algorithm to obtain facial features corresponding to the image information containing the face image of the object to be identified; The facial features are compared with the facial features corresponding to the facial information in the facial information database to draw a judgment conclusion based on the result of the facial feature comparison; wherein the facial information in the facial information database all contains facial feature combinations that have not been split or reorganized.

[0013] Further, for the video picture information, obtaining at least three frames of image picture information that continuously contain the object to be identified, so as to obtain the facial expression changes of the object to be identified in the at least three frames of image picture information that continuously contain the object to be identified; Determining differences in facial features of the object to be identified in the at least three consecutive frames of image information containing the object to be identified based on the aforementioned facial expression changes; The aforementioned facial feature differences are compared with the facial features corresponding to the facial information in the facial information database to draw a judgment conclusion based on the results of the facial feature comparison.

[0014] Furthermore, the comparison is achieved by comparing the similarity between facial features or using a classifier; wherein, the comparison includes extracting at least one facial feature of the aforementioned object to be identified and comparing it with at least one facial feature corresponding to at least one facial information in the aforementioned facial information database.

[0015] An AI recognition device for splitting and recombining facial images, including the following structure: An information acquisition unit is used to acquire picture information containing a face image of a subject to be identified; the picture information includes video picture information and image picture information; the video picture information is composed of multiple frames of continuous image picture information; an information determination unit configured to extract facial information of a subject to be identified from at least one frame of the aforementioned image information, and determine through AI recognition whether the aforementioned facial information contains at least one facial feature that does not belong to the aforementioned subject to be identified; wherein the facial feature that does not belong to the aforementioned subject to be identified refers to at least one facial feature that does not belong to the aforementioned subject to be identified and, after being split and reassembled, becomes at least one facial feature presented in the facial image of the subject to be identified in the aforementioned image information; The information processing unit is configured to, when the judgment is yes, determine the identity information of the person to whom the facial feature actually belongs based on at least one facial feature that does not belong to the aforementioned object to be identified.

[0016] An AI recognition system for facial image splitting and reassembly, comprising: Network nodes, used to send and receive data; An image recognition module is used to identify whether there is a facial image splitting and reassembling behavior in at least one frame of image information containing the facial image of the object to be recognized; A system server, the system server connecting the network node and the image recognition module; The system server is configured to: obtain picture information containing a facial image of an object to be identified; the picture information includes video picture information and image picture information; the video picture information is composed of multiple frames of continuous image picture information; extract the facial information of the object to be identified from at least one frame of the aforementioned picture information, and determine through AI recognition whether there is at least one facial feature in the aforementioned facial information that does not belong to the aforementioned object to be identified; wherein, the facial feature that does not belong to the aforementioned object to be identified refers to at least one facial feature that does not belong to the aforementioned object to be identified, which, after splitting and reorganization, becomes at least one facial feature presented in the facial image of the object to be identified in the aforementioned picture information; when it is determined to be yes, the identity information of the person to whom the facial feature actually belongs is determined based on the at least one facial feature that does not belong to the aforementioned object to be identified.

[0017] A computer-readable storage medium having a computer program stored therein, wherein the computer program, when executed by a processor, implements the implementation steps of any of the above-mentioned AI recognition methods for splitting and recombining facial images.

[0018] Based on the above advantages and positive effects, the advantages of the present invention are: it can identify the behavior of splitting and reassembling the facial information of the object to be identified, determine the identity information of the person corresponding to the facial features that do not belong to the object to be identified, and learn how the picture information of the object to be identified is obtained by splitting and reassembling the facial features of others, thereby identifying fraudulent behavior performed on the network through the splitting and reassembly of facial images. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flowchart provided for an embodiment of the present invention.

[0020] Figure 2 A schematic structural diagram of a device provided in an embodiment of the present invention.

[0021] Figure 3 A schematic diagram of the structure of a system provided in an embodiment of the present invention.

[0022] Description of reference numerals: Device 200, information acquisition unit 201, information judgment unit 202, information processing unit 203; System 300 , network node 301 , image recognition module 302 , system server 303 . DETAILED DESCRIPTION

[0023] The following is a further detailed description of an AI recognition method, device and application for splitting and recombining facial images disclosed in the present invention in conjunction with the accompanying drawings and specific embodiments. It should be noted that the technical features or combinations of technical features described in the following embodiments should not be considered isolated, and they can be combined with each other to achieve better technical effects. In the drawings of the following embodiments, the same reference numerals appearing in each drawing represent the same features or components, which can be applied to different embodiments. Therefore, once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0024] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not intended to limit the conditions under which the invention can be implemented. Any structural modification, change in proportional relationship, or adjustment of size should fall within the scope of the technical content disclosed in the invention without affecting the efficacy and purpose of the invention. The scope of the preferred embodiments of the present invention includes alternative implementations, in which the functions can be performed in a non-described or discussed order, including performing the functions in a substantially simultaneous manner or in a reverse order according to the functions involved, which should be understood by those skilled in the art of the art to which the embodiments of the present invention belong.

[0025] Technologies, methods, and apparatus known to persons of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such technologies, methods, and apparatus should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values. Example

[0026] See also Figure 1FIG. 1 is a flow chart of the present invention. The implementation step S100 of the method is as follows: S101, obtaining image information containing a face image of an object to be identified.

[0027] The acquisition refers to using a face detection algorithm (such as a Haar cascade classifier, a HOG method, or a deep learning method) to obtain image information containing a face image of a to-be-identified object.

[0028] In this embodiment, the object to be identified specifically refers to whether the facial image to be analyzed is an object that has undergone splitting and reassembling.

[0029] The image information of the face image of the object to be identified refers to the detailed visual data containing individual facial features extracted from the face photo or video.

[0030] The picture information is composed of pixels, and the picture information includes video picture information and image picture information.

[0031] The video picture information is composed of multiple frames of continuous image picture information, and each frame of image picture information corresponds to a scene presented in a continuous time.

[0032] When the aforementioned video information is played continuously at a certain frame rate, it can present a dynamic scene. In the image information, each pixel fixedly represents a specific color value, thus forming a static image.

[0033] The image information includes but is not limited to facial geometric features, appearance features, expression changes, lighting conditions, posture information, environmental information, time information, resolution and quality.

[0034] The facial geometric features refer to the facial geometric structure of the object to be identified, such as the size, shape, position of the facial features, and the distance and angle between them.

[0035] The appearance features refer to facial characteristics that can be directly observed visually, including but not limited to the skin color, skin texture, facial lines, facial features, facial proportions, facial symmetry, wrinkles, creases, scars, moles, spots, facial expressions, makeup and jewelry decorations of the face of the object to be identified.

[0036] The skin color refers to the color and tone of the face of the object to be identified, which can reflect race, genetic characteristics and possible health status.

[0037] The skin texture refers to the texture and details of the facial skin surface of the object to be identified, such as the size and distribution of pores, fine lines and wrinkles, etc.

[0038] The facial lines refer to the contour lines of the face of the object to be identified, such as the outline of the chin, the height of the cheekbones, the straightness or curvature of the nose bridge.

[0039] The facial features refer to the shapes and sizes of the eyes, nose, mouth and ears of the object to be identified.

[0040] The facial proportions refer to the relative positions and sizes of the facial features of the object to be identified, as well as the proportional relationship between these facial features.

[0041] The facial symmetry refers to the symmetry between the two sides of the face of the object to be identified.

[0042] The wrinkles and creases refer to dynamic wrinkles caused by facial expressions of the subject to be identified, as well as static wrinkles formed with aging.

[0043] The scars, moles and spots refer to facial features on the face of the subject to be identified caused by genetics, environmental factors, disease or injury.

[0044] The facial expression is used to reflect the impact of the expression changes of the object to be identified on the facial features and facial muscles.

[0045] The makeup is used to reflect the makeup of the face of the object to be identified, such as eye shadow, eyelashes, lip color, etc.

[0046] The decorative accessories are used to reflect the facial decoration of the object to be identified, such as earrings, glasses, hats, etc.

[0047] The expression changes refer to dynamic changes in the face of the object to be recognized, such as details of the expression such as smiling, frowning, blinking, etc.

[0048] The lighting conditions refer to the direction, intensity and color temperature of light that affect the facial image quality and feature visibility of the object to be identified.

[0049] The posture information refers to the angle of the face of the object to be identified, and the posture information includes looking down, looking up, and side face.

[0050] The environmental information refers to the objects and light conditions in the background of the object to be identified, which may affect the recognizability of the face of the object to be identified.

[0051] The time information includes the timestamp information corresponding to at least one frame of picture information involving the facial image of the object to be identified in a video or dynamic monitoring environment; the timestamp information indicates the specific capture time of the facial image of the object to be identified, and the specific capture time can be accurate to seconds or even milliseconds.

[0052] The existence of the time information is helpful for analyzing the changes in the facial expressions of the object to be identified over time.

[0053] The resolution and quality refer to the clarity and pixel size of the facial image of the object to be identified, as well as any compression or damage that may affect the accuracy of recognition.

[0054] In the practical application of facial recognition technology, image information is used to generate a digital representation of a face for matching, identification, or verification. Machine learning and deep learning algorithms are even used to process and analyze this complex data to improve recognition accuracy and efficiency. To obtain high-quality facial images, good lighting conditions and high-resolution image information are preferred to facilitate subsequent analysis.

[0055] S102, extracting facial information of the object to be identified from at least one frame of the image information, and determining through AI recognition whether there is at least one facial feature in the facial information that does not belong to the object to be identified.

[0056] When extracting the facial information of the object to be identified, the extraction operation is performed only on the facial information containing the facial features of the object to be identified in the aforementioned image information. The advantages of this extraction operation are: One of the advantages is that when only focusing on the facial information of the object to be identified, the extraction operation can more attentively capture and extract facial features related to the object to be identified, which helps to improve the accuracy of identification and reduce the possibility of misidentification.

[0057] The second advantage is that by limiting the scope of the extraction operation, the amount of data processed can be reduced. Therefore, compared to processing all facial information in the entire image or video, extracting only the facial features of the object to be identified can be processed more quickly and improve processing efficiency.

[0058] The third advantage is that since the extraction operation focuses on the facial information of the object to be identified, it can reduce sensitivity to factors such as lighting, posture and expression, and it can improve the stability of facial features, making the recognition system more robust to changes in facial information under different environmental conditions.

[0059] A fourth advantage is that by extracting only the facial information of the subject to be identified, the privacy of others can be reduced when multiple people appear in a single image. Protecting personal privacy is crucial in facial recognition applications. Limiting the scope of the extraction operation helps ensure that only the facial features of the subject to be identified are extracted, without involving the facial information of others.

[0060] In actual operation, the accuracy of facial recognition and facial feature comparison may be affected by many factors, such as image quality, lighting conditions, facial expressions, etc. Therefore, in actual operation, it is necessary to adjust the parameters and optimize the algorithm of AI recognition algorithm according to the specific situation to improve the recognition accuracy.

[0061] Based on this, by extracting only the facial information of the subject to be identified, we can improve the accuracy, efficiency, and robustness of recognition while protecting personal privacy. These advantages make the extraction operation more focused and specialized in processing the facial features of the subject to be identified.

[0062] After extracting the facial information of the object to be identified, in order to analyze whether the facial information of the object to be identified has the possibility of being split and reassembled, it is necessary to determine whether there is at least one facial feature in the facial information of the object to be identified that does not originally belong to the object to be identified. The facial feature that does not originally belong to the object to be identified refers to at least one facial feature that does not belong to the object to be identified, which, after being split and reassembled, becomes at least one facial feature presented in the facial image of the object to be identified in the aforementioned picture information. In addition, the facial features that do not originally belong to the object to be identified also include the facial features corresponding to the object to be identified at different ages, which, after being split and reassembled, become at least one facial feature presented in the facial image of the object to be identified in the aforementioned picture information.

[0063] This means that the facial information of the subject to be identified may contain at least one facial feature that does not belong to the subject to be identified after the facial features of the subject to be identified have been split and reassembled. Furthermore, analyzing whether the facial information of the subject to be identified is obtained through splitting and reassembling primarily involves determining whether the facial information of the subject to be identified contains facial features that do not belong to the subject to be identified.

[0064] Furthermore, in actual operation, the split and reassembled image information may have discontinuous or unnatural splicing boundaries, inconsistent feature distribution, or even discontinuous facial features.

[0065] Based on this, this embodiment preferably adopts AI facial recognition technology and feature extraction technology in the existing technology to achieve the above-mentioned judgment operation through AI recognition.

[0066] The advantages of using AI recognition to achieve the above judgment operation are: One of the advantages is that AI recognition can realize automated facial feature analysis without the need for manual inspection one by one, thereby greatly improving the processing speed and efficiency of facial feature extraction.

[0067] The second advantage is that AI-based facial recognition offers high accuracy and consistency. For example, by selecting an appropriate AI model and training and optimizing it with extensive data, it can identify subtle differences in facial features, allowing for more accurate determination of facial features that are not considered to belong to the subject being identified.

[0068] The third advantage is that AI recognition can be applied to large-scale facial data sets and can be adapted to diverse facial features such as different races, ages, and genders. This enables AI recognition to handle a variety of different facial features and provide wide applicability.

[0069] Fourthly, AI recognition can analyze facial features in real time, providing instant judgment results. This is very beneficial for applications that require quick decision-making or real-time feedback, such as facial recognition systems in security monitoring.

[0070] Specifically, the determination includes step S110: S111 , using a facial feature extraction algorithm (such as local binary pattern, principal component analysis, or deep learning model) to obtain facial features corresponding to the image information containing the facial image of the object to be identified.

[0071] S112: Compare the facial features with facial features corresponding to facial information in a facial information database to draw a judgment conclusion based on the result of the facial feature comparison. Wherein, the facial information in the facial information database all includes facial feature combinations that have not been split or reassembled.

[0072] The comparison is preferably performed by comparing the similarity between facial features or using a classifier (e.g., a support vector machine or a deep learning model). The comparison includes extracting at least one facial feature of the subject to be identified and comparing it with at least one facial feature corresponding to at least one facial information in the facial information database.

[0073] It is worth noting that the similarity between the facial features can be measured by setting a threshold. When the similarity of the facial features to be compared is reflected as a high similarity through the threshold, it means that the difference between the two facial features is small, and it can be considered that the facial information of the object to be identified extracted from the picture information contains facial features that do not belong to the object to be identified.

[0074] Therefore, as one of the preferred implementations of this embodiment, when there is at least one facial feature in the picture information containing the facial image of the object to be identified that does not match at least one facial feature corresponding to at least one facial information of the object to be identified in the facial information database, it is determined that there is a facial feature that does not belong to the object to be identified.

[0075] As an example and not a limitation, taking the case where the object to be identified is a person named Alice, the picture information containing Alice's facial image is first obtained, and then the facial features belonging to Alice in the picture information are extracted and compared with the facial features corresponding to Alice's facial information stored in the facial information database.

[0076] If the comparison finds that at least one facial feature of Alice does not match at least one facial feature corresponding to at least one facial information in the facial information database, it can be determined that there are facial features that do not belong to Alice in the picture information containing Alice's facial image, that is, the judgment conclusion is: there are facial features of someone other than Alice in the picture information of the facial image of the object to be identified.

[0077] It is worth mentioning that the facial information database is used to store and manage facial information, preferably by photographing or scanning the face of an individual through a camera or a dedicated face acquisition device to obtain a facial image.

[0078] In practice, the facial information database preferably stores a dataset of publicly available online facial images, serving as a standard for the image information of the facial image of the subject to be identified. Considering the tendency for facial features of the subject to be identified to be disassembled and reassembled, it is preferable to search the aforementioned facial information database for publicly available online facial images as source material, and then use the facial features of the corresponding individuals from these images to replace the facial features of the subject to be identified. Therefore, if at least one facial feature in the image information containing the facial image of the subject to be identified does not match at least one facial feature corresponding to at least one facial information item of the subject to be identified in the facial information database, it is determined that facial features that do not belong to the subject to be identified are present.

[0079] Preferably, when making the judgment, for the video picture information, the following steps are included: S121 , acquiring at least three frames of image information continuously containing the object to be recognized, so as to obtain facial expression changes of the object to be recognized in the at least three frames of image information continuously containing the object to be recognized.

[0080] S122: Determine, based on the facial expression changes, differences in facial features of the object to be identified in the at least three consecutive frames of image information containing the object to be identified.

[0081] In this embodiment, considering that at least three consecutive frames of video information can reflect changes in the facial expression of the object to be identified, for example, the eyebrow position, eye opening degree and mouth shape of the object to be identified in three consecutive frames of image information have significant changes, it can be inferred that the facial expression of the object to be identified has changed during this period of time, and then the difference in facial features of the object to be identified in these three frames of image information can be inferred.

[0082] S123, comparing the aforementioned facial feature differences with the facial features corresponding to the facial information in the facial information database, and drawing a judgment conclusion based on the result of the facial feature comparison.

[0083] In order to determine whether the facial features of the object to be identified in the video picture information are split and reassembled, the differences in the facial features of the object to be identified in the aforementioned at least three consecutive frames of image picture information containing the object to be identified are compared with the facial features corresponding to the facial information in the facial information database, so as to determine whether there is a split and reassembly situation based on the differences in the facial features of the object to be identified in the video picture information.

[0084] It is worth noting that the facial information stored in the facial information database includes the encoding or representation of facial features. These facial feature encodings can be in the form of numerical values ​​or vectors, thereby representing the unique characteristics of each face.

[0085] Therefore, the facial features of the object to be identified are extracted into the same encoding or representation form and compared with the facial features in the database, so that a judgment conclusion can be drawn based on the result of the facial feature comparison. The comparison operation is the same as the comparison operation mentioned in step S112, so it will not be repeated here.

[0086] S103: When the judgment is yes, the identity information of the person to whom the facial feature actually belongs is determined based on at least one facial feature that does not belong to the aforementioned object to be identified.

[0087] After performing the above operations, if the judgment is yes, at least one facial feature that does not belong to the aforementioned object to be identified can be determined. Preferably, these facial features can be found in at least one facial information in the facial information database.

[0088] Thus, based on at least one facial feature that does not belong to the aforementioned object to be identified, the identity information of the person to whom the facial feature actually belongs is determined, wherein the identity information refers to relevant information used to confirm the identity of the person to whom the facial feature actually belongs, specifically including the individual's name, date of birth, nationality, ID number, etc.

[0089] By using at least one facial feature that does not belong to the aforementioned object to be identified as a benchmark, the identity information of the person to whom the facial feature actually belongs is determined. This helps to analyze the entire process of splitting and reassembling the image information containing the facial image of the object to be identified, thereby facilitating the improvement of the learning and optimization capabilities of the AI ​​algorithm in identifying the facial feature splitting and reassembling behavior mentioned in this embodiment.

[0090] Other technical features are described in the previous embodiments and will not be repeated here.

[0091] See also Figure 2 As shown, the present invention also provides an embodiment, providing an AI recognition device 200 for splitting and recombining facial images, including the structure: The information acquisition unit 201 is used to acquire picture information containing the face image of the object to be identified; the picture information includes video picture information and image picture information; the video picture information is composed of multiple frames of continuous image picture information.

[0092] The information judgment unit 202 is used to extract facial information of the object to be identified from at least one frame of the aforementioned image information, and to determine through AI recognition whether there is at least one facial feature in the aforementioned facial information that does not belong to the aforementioned object to be identified; wherein, the facial feature that does not belong to the aforementioned object to be identified refers to at least one facial feature that does not belong to the aforementioned object to be identified, which, after being split and reorganized, becomes at least one facial feature presented in the facial image of the object to be identified in the aforementioned image information.

[0093] The information processing unit 203 is configured to, when the determination is yes, determine the identity information of the person to whom the facial feature actually belongs based on at least one facial feature that does not belong to the aforementioned object to be identified.

[0094] In addition, see Figure 3 As shown, the present invention also provides an embodiment, providing an AI recognition system 300 for splitting and recombining facial images, including: The network node 301 is used to send and receive data.

[0095] The image recognition module 302 is configured to recognize whether there is any facial image splitting and reassembling in at least one frame of image information containing a facial image of a subject to be recognized.

[0096] The system server 303 is connected to the network node 301 and the image recognition module 302 .

[0097] The system server 303 is configured to: obtain screen information containing a facial image of an object to be identified; the screen information includes video screen information and image screen information; the video screen information is composed of multiple frames of continuous image screen information; extract the facial information of the object to be identified from at least one frame of the aforementioned screen information, and determine through AI recognition whether there is at least one facial feature in the aforementioned facial information that does not belong to the aforementioned object to be identified; wherein, the facial feature that does not belong to the aforementioned object to be identified refers to at least one facial feature that does not belong to the aforementioned object to be identified, which, after splitting and reorganization, becomes at least one facial feature presented in the facial image of the object to be identified in the aforementioned screen information; when it is determined to be yes, the identity information of the person to whom the facial feature actually belongs is determined based on the at least one facial feature that does not belong to the aforementioned object to be identified.

[0098] For other technical features, please refer to the previous embodiments and will not be repeated here.

[0099] In addition, an embodiment of the present invention also provides a computer-readable storage medium on which a program is stored for use in the aforementioned AI recognition device for splitting and recombining facial images. When the program is executed by the processor, it can implement the steps of any of the above-mentioned AI recognition methods for splitting and recombining facial images.

[0100] For other technical features, please refer to the previous embodiments and will not be repeated here.

[0101] In the above description, the components may be selectively and operatively combined in any number within the scope of the intended protection of the present disclosure. In addition, terms such as "include," "encompass," and "have" should be interpreted as inclusive or open-ended rather than exclusive or closed by default, unless expressly defined to the contrary. All technical, technological, or other terms have the meanings understood by those skilled in the art, unless they are defined to the contrary. Common terms found in dictionaries should not be interpreted in an overly idealized or unrealistic manner in the context of the relevant technical documentation, unless expressly defined to that extent by the present disclosure.

[0102] Although example aspects of the present disclosure have been described for illustrative purposes, those skilled in the art will appreciate that the foregoing description is merely a description of preferred embodiments of the present invention and does not limit the scope of the present invention in any way. The scope of the preferred embodiments of the present invention includes alternative implementations in which functions may be performed out of the order in which they appear or are discussed. Any changes or modifications made by those skilled in the art based on the foregoing disclosure are intended to fall within the scope of the claims.

Claims

1. An AI recognition method for facial image splitting and reassembly, characterized in that: Including steps: Acquire image information containing a face image of a subject to be identified; the image information includes video image information and image image information; the video image information is composed of multiple frames of continuous image image information; Extracting facial information of the subject to be identified from at least one frame of the image information, and determining through AI recognition whether the facial information contains at least one facial feature that does not belong to the subject to be identified; wherein the facial feature that does not belong to the subject to be identified refers to at least one facial feature that does not belong to the subject to be identified and, after being split and reassembled, becomes at least one facial feature presented in the facial image of the subject to be identified in the image information; When the judgment is yes, the identity information of the person to whom the facial feature actually belongs is determined based on at least one facial feature that does not belong to the aforementioned object to be identified.

2. The method according to claim 1, characterized in that The image information of the face image of the object to be identified refers to the detailed visual data containing individual facial features extracted from the face photo or video; The image information includes facial geometry, appearance, expression changes, lighting conditions, posture information, environmental information, time information, resolution and quality; The facial geometric features refer to the facial geometric structure presented by the object to be identified; Appearance features refer to facial characteristics that can be directly observed visually, including the skin color, skin texture, facial lines, facial features, facial proportions, facial symmetry, wrinkles, creases, scars, moles, spots, facial expressions, makeup and jewelry of the face of the subject to be identified; The expression change refers to the dynamic change of the face of the object to be identified; The lighting conditions refer to the direction, intensity and color temperature of light that affect the quality of the facial image and the visibility of features of the object to be identified; The posture information refers to the angle of the face of the object to be identified, and the posture information includes looking down, looking up and side face; The environmental information refers to the objects and light conditions in the background of the object to be identified; The time information includes timestamp information corresponding to at least one frame of image information involving the facial image of the object to be identified in a video or dynamic monitoring environment; the timestamp information indicates the specific capture time of the facial image of the object to be identified; The resolution and quality refer to the clarity and pixel size of the facial image of the object to be identified, as well as any compression or damage that may affect the accuracy of recognition.

3. The method according to claim 1, characterized in that The facial features that do not belong to the aforementioned object to be identified also include facial features corresponding to different ages of the object to be identified, which are split and reassembled to become at least one facial feature presented in the face image of the object to be identified in the aforementioned picture information.

4. The method according to claim 1, wherein When extracting the facial information of the object to be identified, the extraction operation is performed only on the facial information containing the facial features of the object to be identified in the aforementioned image information.

5. The method according to claim 1, wherein The judgment comprises the steps of: Using a facial feature extraction algorithm to obtain facial features corresponding to the image information containing the face image of the object to be identified; The facial features are compared with the facial features corresponding to the facial information in the facial information database to draw a judgment conclusion based on the result of the facial feature comparison; wherein the facial information in the facial information database all contains facial feature combinations that have not been split or reorganized.

6. The method according to claim 5, characterized in that For the video picture information, obtaining at least three frames of image picture information that continuously contain the object to be identified, so as to obtain the facial expression changes of the object to be identified in the at least three frames of image picture information that continuously contain the object to be identified; Determining differences in facial features of the object to be identified in the at least three consecutive frames of image information containing the object to be identified based on the aforementioned facial expression changes; The aforementioned facial feature differences are compared with the facial features corresponding to the facial information in the facial information database to draw a judgment conclusion based on the results of the facial feature comparison.

7. The method according to claim 5 or 6, characterized in that The comparison is achieved by comparing the similarity between facial features or using a classifier; wherein, the comparison includes extracting at least one facial feature of the aforementioned object to be identified and comparing it with at least one facial feature corresponding to at least one facial information in the aforementioned facial information database.

8. An AI recognition device for facial image splitting and reassembly according to the method of any one of claims 1 to 7, characterized in that: Including structure: An information acquisition unit is used to acquire picture information containing a face image of a subject to be identified; the picture information includes video picture information and image picture information; the video picture information is composed of multiple frames of continuous image picture information; an information determination unit configured to extract facial information of a subject to be identified from at least one frame of the aforementioned image information, and determine through AI recognition whether the aforementioned facial information contains at least one facial feature that does not belong to the aforementioned subject to be identified; wherein the facial feature that does not belong to the aforementioned subject to be identified refers to at least one facial feature that does not belong to the aforementioned subject to be identified and, after being split and reassembled, becomes at least one facial feature presented in the facial image of the subject to be identified in the aforementioned image information; The information processing unit is configured to, when the judgment is yes, determine the identity information of the person to whom the facial feature actually belongs based on at least one facial feature that does not belong to the aforementioned object to be identified.

9. An AI recognition system for facial image splitting and reassembly according to the method of any one of claims 1 to 7, characterized in that include: Network nodes, used to send and receive data; An image recognition module is used to identify whether there is a facial image splitting and reassembling behavior in at least one frame of image information containing the facial image of the object to be recognized; A system server, the system server connecting the network node and the image recognition module; The system server is configured to: obtain picture information containing a facial image of an object to be identified; the picture information includes video picture information and image picture information; the video picture information is composed of multiple frames of continuous image picture information; extract the facial information of the object to be identified from at least one frame of the aforementioned picture information, and determine through AI recognition whether there is at least one facial feature in the aforementioned facial information that does not belong to the aforementioned object to be identified; wherein, the facial feature that does not belong to the aforementioned object to be identified refers to at least one facial feature that does not belong to the aforementioned object to be identified, which, after splitting and reorganization, becomes at least one facial feature presented in the facial image of the object to be identified in the aforementioned picture information; when it is determined to be yes, the identity information of the person to whom the facial feature actually belongs is determined based on the at least one facial feature that does not belong to the aforementioned object to be identified.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps according to any one of claims 1 to 7 are implemented.

Images