User management method and device based on smart pension

Abstract

The invention is suitable for the technical field of face recognition, and provides a user management method and device based on smart pension, and the method comprises the steps: collecting a to-be-recognized face image through a camera device, and obtaining a plurality of pre-stored standard face features; extracting real-time face features of the face image to be recognized; the real-time face features comprise five sense organ distribution features and real-time wrinkle texture features; calculating a vector similarity between the standard face feature and the real-time face feature; when the vector similarity is greater than a preset value, taking an activity area corresponding to the camera device as user management information; wherein the user management information is used for recording the daily activity type of the user. According to the user management method based on smart pension, through an improved face recognition technology and real-time user activity management, the recognition accuracy, stability and management efficiency of the system are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the technical field of facial recognition, and in particular relates to a user management method and device based on smart elderly care. Background Technology

[0002] With the accelerating aging of society, smart elderly care has become an important means of addressing the care needs of the elderly. Smart elderly care utilizes modern information technology and intelligent devices to provide comprehensive elderly care services, improving the quality of life and safety of seniors. User management is a crucial component of a smart elderly care system; efficiently and accurately managing and recording the daily activities of elderly users is of great significance for improving service quality and response speed.

[0003] Traditional user management methods typically rely on manual recording or simple electronic devices, such as manually filled-out forms or simple sensors. These methods are not only inefficient but also prone to errors, failing to meet the demands for real-time performance and accuracy. With the development of facial recognition and intelligent image processing technologies, camera-based user management methods are gaining increasing attention. These methods, by capturing users' facial images and combining them with facial recognition technology, achieve automated and real-time user management.

[0004] Existing facial recognition technologies primarily rely on the comparison of standard facial features, which are typically static and two-dimensional, such as the position and shape of facial features like the eyes, nose, and mouth. However, as people age, their facial features change, such as increased wrinkles and skin laxity, which affects the accuracy of traditional facial recognition algorithms. Therefore, relying solely on static facial features for facial recognition is insufficient to meet the high-precision user management requirements of smart elderly care systems. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a user management method and device based on smart elderly care, in order to solve the technical problem that relying solely on static facial features for face recognition is insufficient to meet the high-precision user management requirements of smart elderly care systems.

[0006] A first aspect of this invention provides a user management method based on smart elderly care, the user management method based on smart elderly care comprising: The camera device captures images of the faces to be identified and obtains multiple pre-stored standard facial features; Extract real-time facial features from the face image to be identified; the real-time facial features include facial feature distribution features and real-time wrinkle texture features; Calculate the vector similarity between the standard facial features and the real-time facial features; When the vector similarity is greater than a preset value, the activity area corresponding to the camera device is used as user management information; wherein, the user management information is used to record the user's daily activity type.

[0007] Furthermore, the step of extracting real-time facial features from the face image to be identified includes: The facial contours in the image of the person to be identified are extracted based on the Haar cascade classifier; The left corner of the eye, the right corner of the eye, and the tip of the nose in the image of the face to be identified are extracted using a facial landmark detector. Based on the distribution characteristics of the facial contour, the position of the left corner of the eye, the position of the right corner of the eye, and the position of the tip of the nose, the facial feature distribution characteristics are constructed. Extract a local facial region located between the corner of the eye and the chin from the face region, and remove the preset nose region from the local facial region to obtain the cheek region; wherein, the corner of the eye refers to the straight line formed by the left corner of the eye and the right corner of the eye; Extract real-time wrinkle texture features from the cheek region; The facial feature distribution features and the real-time wrinkle texture features are used as the real-time face features.

[0008] Furthermore, the step of constructing the facial feature distribution features based on the distribution characteristics between the facial contour, the position of the left corner of the eye, the position of the right corner of the eye, and the position of the tip of the nose includes: Calculate the first distance between the tip of the nose and the left corner of the eye, and calculate the second distance between the tip of the nose and the right corner of the eye; Calculate the average distance between the first distance and the second distance; A circular region is constructed with the nose tip position as the center and the average distance as the radius; Extract the intersection point between the circular region and the facial contour; Calculate the third distance between the intersection point and the positions of the left corner of the eye, the right corner of the eye, and the tip of the nose, respectively; Calculate the fourth distance between the positions of the left and right corners of the eyes; The first distance, the second distance, multiple third distances, and the fourth distance are normalized and combined in a fixed order to form the facial feature distribution.

[0009] Furthermore, the step of extracting real-time wrinkle texture features of the cheek region includes: The cheek area is processed to obtain a binarized image; Extract the texture region corresponding to the maximum value in the binarized image; For each texture region, extract the first pixel located at the leftmost edge of the texture region, extract the second pixel located at the rightmost edge of the texture region, and extract the third pixel located at the center of the texture region; The first pixel, the second pixel, and the third pixel are designated as marker pixels. For each marker pixel, extract the pixel neighborhood of a preset size around that marker pixel; Extract target texture units from the pixel neighborhood; wherein, the target texture unit is a set of pixels with consistent features; Real-time wrinkle texture features are extracted based on the target texture unit.

[0010] Further, the step of extracting the target texture unit in the pixel region includes: Based on pixel neighborhood, directional response intensity and dominant polarity are extracted in multiple preset directions; wherein, the multiple preset directions include 0°, 45°, 90° and 135°; the directional response intensity is used to quantify the significance of structural changes in the pixel neighborhood in the preset directions; the dominant polarity is used to quantify the brightness transition trend in the pixel neighborhood in the preset directions. The pixel corresponding to the maximum directional response intensity is used as the seed point, and the neighboring pixels around the seed point are obtained. When the first adjacent pixel of the seed point meets the preset conditions, the seed point and the first adjacent pixel are merged into a texture unit; wherein, the preset conditions include that the dominant polarity of the two is the same and the difference in directional response intensity between the two is less than a threshold. When the second adjacent pixel of a texture unit meets the preset conditions, the texture unit and the second adjacent pixel are merged into a new texture unit; The step of merging the texture unit and the second adjacent pixel into a new texture unit when the second adjacent pixel of the texture unit meets the preset conditions is executed iteratively to obtain the target texture unit.

[0011] Furthermore, the step of extracting real-time wrinkle texture features based on the target texture unit includes: Calculate the first perimeter of the target texture unit; Construct a circle with the same area as the target texture unit, and extract the second circumference of the circle; Subtracting the first perimeter from the second perimeter and dividing by the second perimeter yields the boundary irregularity; wherein the boundary irregularity is used to measure the degree of boundary irregularity. The statistical analysis identifies the enclosed hole regions within the target texture unit; where a hole region refers to an independent region enclosed by the target texture unit. Divide the number of pixels in all hole regions by the total number of pixels in the target texture unit to obtain the percentage of hole regions; In the binarized image, for each current pixel in the target texture unit, obtain the eight adjacent pixels of that current pixel; When all eight pixels are pixels in the target texture unit, the current pixel is taken as a connected component; Divide the number of connected components by the total number of pixels in the target texture unit to obtain the proportion of connected components; Extract the center point of the texture region, calculate the first angle between the center point and the left corner of the eye, and calculate the second angle between the center point and the right corner of the eye; The boundary irregularity, the proportion of the hole region, the proportion of the connected components, the first included angle, and the second included angle are used as real-time wrinkle texture features.

[0012] Furthermore, the standard facial features include a first standard facial feature and a second standard facial feature; The step of calculating the vector similarity between the standard facial features and the real-time facial features includes: Calculate the first similarity between the first standard facial features and the facial feature distribution features; Calculate the second similarity between the second standard face features and the real-time wrinkle texture features; If both the first similarity and the second similarity are greater than or equal to a preset similarity, then the minimum value between the first similarity and the second similarity is taken as the vector similarity. If the first similarity or the second similarity is less than the preset similarity, then the first similarity and the second similarity are weighted and summed to obtain the vector similarity. If both the first similarity and the second similarity are less than the preset similarity, then the vector similarity is 0.

[0013] A second aspect of the present invention provides a user management device based on smart elderly care, comprising: The acquisition unit is used to acquire images of the face to be identified through a camera device and obtain multiple pre-stored standard facial features; An extraction unit is used to extract real-time facial features from the face image to be identified; the real-time facial features include facial feature distribution features and real-time wrinkle texture features; A calculation unit is used to calculate the vector similarity between the standard facial features and the real-time facial features; The judgment unit is used to determine the activity area corresponding to the camera device as user management information when the vector similarity is greater than a preset value; wherein the user management information is used to record the user's daily activity type.

[0014] A third aspect of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps in the user management method based on smart elderly care described in the first aspect.

[0015] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the user management method based on smart elderly care described in the first aspect.

[0016] The beneficial effects of this invention compared to existing technologies are as follows: by combining facial feature distribution characteristics and real-time wrinkle texture characteristics, the real-time features of the face image to be identified are comprehensively extracted. Compared to traditional methods that rely solely on static facial features, this invention can more accurately capture subtle facial features that change over time in the elderly, thereby improving the accuracy of face recognition, especially when facial features change significantly in the elderly. By calculating the vector similarity between standard facial features and real-time facial features, and making judgments based on a preset similarity threshold, this invention ensures the reliability of user identification. Even under different lighting conditions and changes in facial expressions, the system can still perform face recognition stably and accurately, ensuring the reliability and effectiveness of user management information. The activity area corresponding to the camera device is used as user management information to record the user's daily activity types in real time. This real-time user activity management function not only improves the timeliness of data but also provides accurate user behavior data support for other service modules in the smart elderly care system, which helps to implement personalized service and care plans. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic flowchart of a user management method based on smart elderly care provided by the present invention is shown; Figure 2 This diagram illustrates a user management device based on smart elderly care according to an embodiment of the present invention. Figure 3 A schematic diagram of a terminal device provided in an embodiment of the present invention is shown. Detailed Implementation

[0019] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0020] This invention provides a user management method and device based on smart elderly care, which solves the technical problem that relying solely on static facial features for face recognition is insufficient to meet the high-precision user management requirements of smart elderly care systems.

[0021] First, this invention provides a user management method based on smart elderly care. Please see [link / reference]. Figure 1 , Figure 1 This diagram illustrates a schematic flowchart of a user management method based on smart elderly care provided by the present invention. Figure 1 As shown, this user management method based on smart elderly care may include the following steps: Step 101: Acquire the image of the face to be identified using a camera device and obtain multiple pre-stored standard facial features; Cameras deployed in senior living facilities are used to capture real-time facial images of users requiring identification. The system retrieves standard facial feature templates from a pre-built database of registered users. This database is either created during user registration or updated periodically.

[0022] Step 102: Extract the real-time facial features of the face image to be identified; the real-time facial features include facial feature distribution features and real-time wrinkle texture features; Feature extraction is performed on the real-time face image acquired in step 101. First, spatial relationship information such as the position, size, relative distance, and angle of key facial features like the eyes, nose, mouth, and eyebrows is extracted. Second, the system extracts subtle skin texture information such as the depth, direction, density, length, and distribution pattern of wrinkles in specific facial areas (e.g., the corners of the eyes, forehead, and nasolabial folds).

[0023] Wrinkle patterns exhibit strong individual variability and relative stability (changing slowly over time), representing a more refined biometric feature than the macroscopic distribution of facial features, effectively distinguishing similar-looking facial features. While partial occlusion (such as wearing a mask covering the mouth and nose) or facial expressions significantly impact facial feature distribution, wrinkles on the forehead and corners of the eyes may still retain high visibility and recognition value. In elderly care scenarios, a user's appearance changes with age, but the patterns of wrinkle changes themselves contain identity information. Standard wrinkle features recorded during registration can be updated as the user updates.

[0024] Combined with the relatively static distribution of facial features, it forms a dual recognition factor of "macro layout + micro texture", which greatly improves the overall recognition accuracy and robustness.

[0025] Specifically, step 102 includes steps 1021 to 1026: Step 1021: Extract the facial contours from the face image to be identified based on the Haar cascade classifier; This method quickly and accurately outlines the approximate rectangular region (i.e., facial contour) containing a face in an image. This provides a basic scope for subsequent, more refined feature point localization and region extraction, avoiding blindly searching for key points throughout the entire image and improving efficiency and accuracy. The Haar cascade classifier is a traditional technique and will not be considered in any way here.

[0026] Step 1022: Extract the positions of the left corner of the eye, the right corner of the eye, and the tip of the nose from the face image to be identified using a facial landmark detector; Use a facial landmark detector (which can be a traditional Dlib, MediaPipe, or a deep learning-based model). This type of algorithm can accurately predict the coordinates of a series of predefined points on a face.

[0027] The left corner of the eye (left outer canthus), the right corner of the eye (right outer canthus), and the tip of the nose are the core reference points for constructing the "distribution features of facial features" (see next step), and also the geometric basis for subsequently defining the "eye corner line" and locating the "cheek region." The accuracy of the location directly affects the effect of all subsequent steps.

[0028] Step 1023: Construct the facial feature distribution features based on the distribution characteristics between the facial contour, the position of the left eye corner, the position of the right eye corner, and the position of the nose tip; These calculated geometric parameters (such as distance, angle, and proportion) are combined into a feature vector, representing the approximate spatial layout of the eyes and nose on this face. This is the numerical representation of the "facial feature distribution characteristics".

[0029] Specifically, step 1023 includes steps A1 to A7: Step A1: Calculate the first distance between the tip of the nose and the left corner of the eye, and calculate the second distance between the tip of the nose and the right corner of the eye; Obtain key distance information between the nose and eyes. The first and second distances directly reflect the length of the bridge of the nose and the relative position of the nose and eyes.

[0030] Step A2: Calculate the average distance between the first distance and the second distance; The average distance between the left and right sides is taken to reduce the impact of unilateral asymmetry caused by slight head tilt or facial expression, thereby improving feature stability.

[0031] The average distance will serve as the baseline radius for defining key regions in subsequent steps. It represents the "average" distance from the nose to the eyes in a current face, and is a scale information related to the size of an individual's face.

[0032] Step A3: Construct a circular region with the nose tip position as the center and the average distance as the radius; Create a circle centered on the nose with a radius equal to the "average nose-to-eye distance". The size of this circle adapts to different faces (larger faces have larger average distances, so the circle is larger; smaller faces have smaller average distances, so the circle is smaller), providing a relatively normalized spatial framework for subsequent feature extraction. This region aims to include facial structures where the nose and eyes are closely related.

[0033] Step A4: Extract the intersection point between the circular region and the facial contour; Perform geometric intersection operations to find all intersection points between the circle defined in step A3 and the facial contour boundary obtained through Haar cascade detection.

[0034] These intersections concretize the spatial relationship between the circle and the facial contour. These points represent the specific contact positions of a circle centered on the tip of the nose and with the average distance between the nose and eyes as its radius on the facial boundary, given the current size and shape of the face. They contain information about the width and contour shape of the face (such as square face or round face) relative to the position of the nose and eyes.

[0035] Step A5: Calculate the third distance between the intersection point and the positions of the left corner of the eye, the right corner of the eye, and the tip of the nose, respectively; For each intersection point obtained in the previous step, calculate the distance from that intersection point to the left corner of the eye. Calculate the distance from that intersection point to the right corner of the eye. Calculate the distance from that intersection point to the tip of the nose.

[0036] This step establishes spatial relationships between specific points on the facial contour boundary and core facial feature points. This set of distances quantifies the spatial distribution pattern of the facial contour relative to the core facial features and is the core data describing the "distribution characteristics." The existence of intersection points allows the measurement to go beyond the core points themselves, introducing contour constraints.

[0037] Step A6: Calculate the fourth distance between the left and right corners of the eye; Step A7: Normalize the first distance, the second distance, multiple third distances and the fourth distance, and combine them in a fixed order to form facial feature distribution characteristics.

[0038] A large face close to the camera and a small face far from the camera have vastly different original distance values, but their proportions after normalization should be similar. This is key to ensuring feature scale invariance, making the recognition algorithm robust to faces captured at different distances.

[0039] All normalized distance values ​​are arranged in a predefined, fixed order to form a one-dimensional feature vector. Normalization ensures the comparability of features; combining them in a fixed order forms a structured, numerical feature vector representing the spatial distribution of facial features of a specific face. This is the final constructed "facial feature distribution feature." Each component represents a relative measure of the face shape and the positional relationship of facial features.

[0040] In the embodiments corresponding to steps A1 to A7, the method is achieved entirely through precise point coordinate distance calculation and geometric region intersection, making it objective, calculable, and repeatable. The key innovation lies in using the intersection of a circular region and the facial contour as a bridge to dynamically link core feature points with the facial boundary. This contains richer information about the facial contour and relative positions of facial features than simply measuring the distance between a few core points. A circle is constructed with the tip of the nose as the center and the average nose-to-eye distance as the radius, allowing the reference region to adapt to different individuals' face sizes and proportions, providing a personalized, scale-dependent spatial framework for subsequent intersection point extraction and distance measurement. Measurements cover various distance relationships between core points (nose-left eye, nose-right eye, left eye-right eye) and between core points and contour constraint points, providing rich information. Normalizing all distance measurements (dividing by the average distance or inter-eye distance) completely eliminates the influence of the absolute size (distance) of the face in the image, giving the constructed facial feature distribution features strong scale invariance. This is a key design feature to ensure the algorithm's robustness to different shooting distances (users moving closer to or further from the camera) in elderly care environments. All normalized distance values ​​are combined into a vector in a fixed order to ensure consistency in feature representation. The entire scheme comprehensively utilizes the relative spatial information of eye and nose positions and facial contour shapes. The generated normalized feature vectors can effectively distinguish the facial feature distribution patterns of different individuals, while also exhibiting a certain tolerance to scale variations (camera distance) and slight changes in head pose (due to the use of average distance and contour constraint points).

[0041] Step 1024: Extract a local facial region located between the eye corner line and the chin in the face region, and remove the preset nose region in the local facial region to obtain the cheek region; wherein, the eye corner line refers to the straight line formed by the left eye corner position and the right eye corner position; Within the entire detected face area, a local region is extracted. The upper boundary of this region is the "eye corner line". Its lower boundary is the bottom of the chin (the chin position can be obtained by facial contour or facial key point detection).

[0042] Within this area extending from the outer corner of the eye to the chin, the nose exists (the tip of the nose is already located). Nasal region textures (such as the bridge and wings of the nose) are typically not typical wrinkle areas (nasolabial folds are on either side of the wings). Therefore, a pre-defined "pre-defined nose region" (such as a rectangular or oval area of ​​appropriate size centered on the tip of the nose) needs to be eliminated.

[0043] After removing the nose area, the remaining main parts are the left and right cheek areas (including the nasolabial folds, part of the cheekbone, and the area near the corners of the mouth). This area is rich in wrinkles (especially nasolabial folds and perioral wrinkles).

[0044] The cheek area is a common location for deep wrinkles (nasolabial folds), and the texture is relatively controllable by facial expressions (compared to forehead and crow's feet).

[0045] By using relatively stable and precisely located points such as the corner of the eye and the chin to define the boundaries of the region, the stability and consistency of the region positioning are improved (even if the head rotates or tilts slightly).

[0046] Actively removing the nose as a non-target texture area avoids interference from non-wrinkle information such as nose shape and gloss on texture feature extraction, thus improving the purity and representativeness of subsequent wrinkle feature extraction.

[0047] The preset nose area is based on the distribution of facial contours, and the nose distribution position is preset.

[0048] Step 1025: Extract the real-time wrinkle texture features of the cheek area; The skin wrinkle patterns (deep grooves, shallow lines, direction, density, etc.) in the cheek area are transformed into a numerical feature vector, namely "real-time wrinkle texture features".

[0049] Specifically, step 1025 includes steps B1 to B7: Step B1: Perform grayscale processing on the cheek area to obtain a binarized image; The cheek area (RGB color image) is converted to a grayscale image. This reduces the amount of data, simplifies subsequent processing, and wrinkle information is primarily contained in brightness variations. Binarization thresholding is applied to the grayscale image. A threshold is set to convert the brightness value of each pixel in the grayscale image to pure black (0) or pure white (255). Wrinkles typically appear as grooves darker than the surrounding skin (under appropriate lighting and thresholding). A clear black-and-white binary image is generated. In this image, the target (wrinkles or skin texture grooves) is typically highlighted as black areas (or white areas, depending on the threshold setting), while the background (flat skin) is displayed in the opposite color. This greatly simplifies the localization and shape analysis of wrinkle areas.

[0050] Step B2: Extract the texture region corresponding to the maximum value in the binarized image; The binary image is traversed, and the attributes of each connected region (composed of adjacent pixels of the same color) are labeled and calculated. Then, the connected region with the highest value is selected (the highest value indicates the presence of wrinkle shadows). This step aims to locate and extract the most prominent wrinkle structure regions in the image, which will be the core target for subsequent analysis.

[0051] Step B3: For each texture region, extract the first pixel located on the leftmost side of the texture region, extract the second pixel located on the rightmost side of the texture region, and extract the third pixel located at the center of the texture region; Step B4: Designate the first pixel, the second pixel, and the third pixel as marker pixels; These three points (the first, second, and third pixels) are designed to cover the key spatial locations of the target wrinkle region: the two endpoints and the center point. They delineate the approximate horizontal span (length) and center location of the wrinkle, providing anchor points for subsequent precise sampling of local texture features within its neighborhood. This approach is more representative than random sampling.

[0052] Step B5: For each marker pixel, extract the pixel neighborhood of a preset size around that marker pixel; Centered on each of the marker pixels (first pixel, second pixel, and third pixel) determined in step B4, extract a square or rectangular image block of fixed size (preset size) from the surrounding image. This image block is called the pixel neighborhood of that marker pixel.

[0053] The preset pixel neighborhood size is an N x N pixel window (e.g., 50x50, 100x100, 200x200). If the neighborhood is too small, there will be insufficient information; if the neighborhood is too large, it will contain too much irrelevant information or cross-wrinkle structures. It needs to be set according to the actual calculation requirements.

[0054] The focus of the analysis is shifted from a single point to a local region containing that point and its immediate surroundings. The texture patterns (direction, contrast, pattern) of wrinkles need to be effectively represented within a local window.

[0055] Step B6: Extract the target texture unit in the pixel neighborhood; wherein the target texture unit is a set of pixels with consistent features; Within the pixel neighborhood (small image patch) corresponding to each marker point, a sub-region, called the "target texture unit," is further analyzed and located. This unit is defined as a set of pixels with "feature consistency" within its neighborhood. This avoids averaging or generalizing the entire neighborhood window. The core "units" that best represent the local structural features of wrinkles (such as a segment of a wrinkle groove or the area around a wrinkle intersection) within the neighborhood are precisely located and extracted. This improves the discriminative power of the features and the sensitivity to subtle changes.

[0056] Specifically, step B6 includes steps B61 to B65: Step B61: Based on the pixel neighborhood, extract the directional response intensity and dominant polarity in multiple preset directions respectively; wherein, the multiple preset directions include 0°, 45°, 90° and 135°; the directional response intensity is used to quantify the significance of structural changes in the pixel neighborhood in the preset directions; the dominant polarity is used to quantify the brightness transition trend in the pixel neighborhood in the preset directions. Multiple preset directions include, but are not limited to, 0°, 45°, 90°, and 135°. The four directions of 0° (horizontal), 45° (diagonal), 90° (vertical), and 135° (anti-diagonal) cover the basic texture direction.

[0057] 1. The calculation logic for directional response intensity is as follows: ① Define an asymmetric sector region: Centered on the marker pixel, define two points about the preset direction. Asymmetrical sector region: Positive sector The coverage angle range is approximately as follows: Or 30°).

[0058] negative sector The coverage angle range is approximately as follows: . Notice: yes The opposite direction.

[0059] These two sector areas are physically symmetrical (not axially symmetrical) with respect to the marker pixel, but they surround a preset direction. The positive and negative sides. They do not overlap and only cover a specific angular range, not the entire circle.

[0060] ② Calculate the weighted gradient contribution (asymmetric response): for the sector region Each pixel (i, j) within: Calculate the spatial offset vector of this point relative to the marker pixel (x, y). .

[0061] Calculate the grayscale gradient magnitude at this point. (The Sobel or Prewitt operator can be used).

[0062] Calculate the gradient direction at this point and Directional consistency weight .in It is the gradient direction at point (i, j). In the same direction When it is close to 1, it is close to 0 when it is perpendicular.

[0063] Calculate spatial distance weights The closer to the center, the greater the weight. This is the attenuation factor, used to control the attenuation rate.

[0064] The positive contribution of this point is: .

[0065] Similarly, for the sector region Each pixel (i, j) within: Calculate negative contribution The directional consistency weight is calculated as follows: Spatial distance weight Calculation method and Same (based on the same distance). Note the direction. .

[0066] The above calculation process is used to aggregate the contribution of pixels within the positive and negative sector regions to the directional trend, respectively.

[0067] ③ Aggregation direction response: Calculate the positive aggregation response: ; Calculate the negative aggregation response: ; Directional response intensity This value represents the direction. Above, the overall intensity of local structural changes. ; in, These are sector areas The number of effective pixels within the area (used for normalization to avoid bias caused by differences in region size). K is an optional scaling constant.

[0068] 2. The calculation logic for dominant polarity is as follows: Dominant polarity Describes the direction Above, the dominant trend is the change in intensity.

[0069] Calculate the net polarity difference: ; Set threshold (a small positive number, for example) , (A constant used to prevent division by zero): if :but (Indicates positive dominance, for example in) The direction shows a transitional trend from dark to light.

[0070] if :but (Indicates negative dominance, for example in) The direction shows a transition trend from light to dark.

[0071] The above calculation method effectively captures the intensity change patterns and trends in a specific direction within the pixel neighborhood, providing richer, more directional, and polarity-aware basic information for subsequent similarity-guided region growth. This is a key innovation that distinguishes it from traditional gradient calculation.

[0072] This calculation method is asymmetric, and its core lies in using two asymmetric, point-symmetric sector regions. Capture direction separately The changes on both the positive and negative sides. This is different from the symmetric gradient operator.

[0073] This calculation method employs weighted aggregation, combining the gradient magnitude G and directional consistency weights in the response calculation. (Emphasis on the direction of the goal) or +180 consistency) and spatial distance weighting (Emphasis on pixels near the center).

[0074] Directional response intensity The sum of the responses in the negative sector region determines the total intensity of change in that direction.

[0075] Dominant polarity It is determined by the sum of the responses of the negative sector region, reflecting the total intensity of local change in that direction.

[0076] Dominant polarity The difference in response between the positive and negative sector regions reveals the dominant trend of intensity change in that direction ("+" from the center along...). (Observe the color that brightens, and the color that darkens).

[0077] For each marker pixel (x, y), a value is calculated for each preset direction (0°, 45°, 90°, 135°). .

[0078] Step B62: Use the pixel corresponding to the maximum directional response intensity as the seed point, and obtain the neighboring pixels around the seed point; The pixel with the strongest directional response within its neighborhood is selected as the seed point to ensure that growth begins with the most significant texture feature. Its 8-neighbor or 4-neighbor pixels are then selected as candidate growth points.

[0079] Step B63: When the first adjacent pixel of the seed point meets the preset conditions, merge the seed point and the first adjacent pixel into a texture unit; wherein, the preset conditions include that the dominant polarity of the two is the same and the difference in directional response intensity between the two is less than a threshold. Generate initial texture units (composed of a seed point and its adjacent points that meet certain conditions). Core decision logic: Condition 1: Same dominant polarity → ensures that merged pixels belong to the same texture structure (e.g., all are edges of wrinkles and depressions). Condition 2: Intensity difference < threshold → avoids crossing discontinuous regions with abrupt changes in intensity (e.g., the boundary between a mole and a wrinkle).

[0080] Step B64: When the second adjacent pixel of the texture unit meets the preset conditions, merge the texture unit with the second adjacent pixel into a new texture unit; The growth target changes from a single seed point to the entire current texture unit. It examines all neighboring pixels surrounding the unit (not just the neighborhood of the original seed point).

[0081] Step B65: Iteratively execute the step of merging the texture unit and the second adjacent pixel into a new texture unit when the second adjacent pixel of the texture unit meets the preset conditions, to obtain the target texture unit.

[0082] The merging operation is performed iteratively until no new pixels can be merged, resulting in the target texture unit. The termination condition is that no neighboring pixels satisfy the merging condition (same polarity + intensity difference < threshold). Finally, a connected, homogeneous texture unit (i.e., the target texture unit) is generated. A complete microtexture structure is generated, preserving the continuity of wrinkles.

[0083] In the embodiments corresponding to steps B61 to B65, the directional response intensity (edge ​​saliency) + dominant polarity (structural concavity / convexity) approach closely approximates the biological essence of wrinkles. Dual-condition constraints (polarity + intensity difference) effectively resist noise such as uneven illumination and age spots. The intensity difference threshold dynamically accommodates both shallow wrinkles (weak edges) and deep wrinkles (strong edges). This scheme achieves precise segmentation of wrinkle microstructures through a triple joint criterion of direction, polarity, and intensity. Step B7: Extract real-time wrinkle texture features based on the target texture unit.

[0084] The features (descriptors of target texture units) extracted from three key locations (the start / left end, middle, and end / right end of the wrinkle) that best represent the local wrinkle structure are combined into a complete, numerical feature representation. This vector captures the microscopic texture characteristics of the target wrinkle at key spatial points.

[0085] In the embodiments corresponding to steps B1 to B7, high-discrimination wrinkle texture features are extracted from the image of the cheek region. The core of this method lies in locating key points, defining local neighborhoods, and refining and extracting the most consistent core "texture units," ultimately generating features based on these units. This method particularly emphasizes capturing the local microstructure of wrinkles, which is one of the key technical details supporting high-precision recognition.

[0086] The real-time wrinkle texture feature is composed of the wrinkle texture features corresponding to the first pixel, the second pixel, and the third pixel.

[0087] Specifically, step B7 includes steps B71 to B710: Step B71: Calculate the first perimeter of the target texture unit; The number of contour pixels (chain code length) or the accumulated Euclidean distance are counted using boundary tracing algorithms (such as Moore's neighborhood tracing). The first perimeter reflects the edge complexity of the wrinkle texture (deeper wrinkles have more tortuous boundaries).

[0088] Step B72: Construct a circle with the same area as the target texture unit, and extract the second circumference of the circle; Calculate the area S (number of pixels) of the texture unit. Construct an ideal circle with the same area S (radius r = √(S / π)). Calculate the circumference of this circle P2 = 2πr.

[0089] The second perimeter provides a boundary reference for a smooth, continuous shape.

[0090] Step B73: Subtract the first perimeter from the second perimeter and divide by the second perimeter to obtain the boundary irregularity; wherein, the boundary irregularity is used to measure the degree of boundary irregularity; Boundary irregularity measures the fractal properties of texture boundaries. Fractal geometry theory is applied to the assessment of microwrinkles, showing sensitivity to boundary roughness caused by age-related skin laxity.

[0091] Step B74: Count the hole regions surrounded by the target texture unit; where a hole region refers to an independent region surrounded by the target texture unit. Identify enclosed background regions (holes) within texture units.

[0092] Step B75: Divide the number of pixels in all hole regions by the total number of pixels in the target texture unit to obtain the percentage of hole regions; Calculate the total number of pixels in the hole region and divide it by the total number of pixels in the target texture unit to obtain the proportion of the hole region. This proportion reflects the density of holes inside the texture unit.

[0093] The percentage of the area with holes is used to capture the structural breakage features inside the texture unit. High percentage → more wrinkle breaks (such as age spots / pore interference), low percentage → continuous texture (deep straight wrinkles).

[0094] Step B76: In the binarized image, for each current pixel in the target texture unit, obtain the eight adjacent pixels of that current pixel; Step B77: When all eight pixels are pixels in the target texture unit, the current pixel is taken as a connected component; Step B78: Divide the number of connected components by the total number of pixels in the target texture unit to obtain the proportion of connected components; The total number of connected components is calculated and divided by the total number of pixels in the target texture unit to obtain the proportion of connected components. This proportion reflects the connectivity and density of the target texture unit. High proportion → fragmentation (shallow fine lines), low proportion → continuous structure (deep wrinkles).

[0095] Step B79: Extract the center point of the texture region, calculate the first angle between the center point and the left corner of the eye, and calculate the second angle between the center point and the right corner of the eye; Determine the geometric center point of the textured region and calculate the angles between this center point and the left and right corners of the eye. The first angle is the angle between the center point and the left corner of the eye, and the second angle is the angle between the center point and the right corner of the eye. These angles provide information about the location and direction of the wrinkles. Mapping the local texture to the global face coordinate system enhances pose robustness (relative angles remain stable even when the head is turned).

[0096] Step B710: Use the boundary irregularity, the proportion of the hole region, the proportion of the connected components, the first included angle and the second included angle as real-time wrinkle texture features.

[0097] The five metrics (boundary irregularity, proportion of hole region, proportion of connected components, first angle, and second angle) corresponding to the first, second, and third pixels obtained from the above calculations are used as real-time wrinkle texture features. These features can comprehensively reflect the shape, density, and spatial distribution of wrinkles on the face, providing rich information for subsequent analysis and processing.

[0098] In the embodiments corresponding to steps B71 to B710, a specific method for extracting real-time wrinkle texture features from the target texture unit is described. By calculating boundary irregularity, the proportion of hole regions, the proportion of connected components, and the geometric position angle of the texture region, a series of feature values ​​are obtained. These features can be used for fine analysis of facial wrinkles in a smart elderly care system, which helps improve the system's ability to understand and monitor users' facial expressions and health status.

[0099] Step 1026: Use the facial feature distribution features and the real-time wrinkle texture features as the real-time face features.

[0100] This process generates a complete "real-time facial feature" representation of the face image to be identified, encompassing both macroscopic facial feature layout information and microscopic cheek wrinkle texture information. This comprehensive feature vector will be used for subsequent similarity calculations with standard features in the database.

[0101] In the embodiments corresponding to steps 1021 to 1026, key points (corners of the eyes, tip of the nose) are located and combined with facial contours. Geometric relationships (distance, position, proportion) are used to quantify the macroscopic distribution, making the method clear and specific. The lower boundary is defined using the line at the corner of the eye and the chin. A preset nose area is actively eliminated. Finally, a texture analysis algorithm is applied to the cheek area. Key point detection ensures basic positioning accuracy. Geometrically constrained region definition reduces the impact of pose and angle changes. Eliminating the nose improves the representativeness of wrinkle features. Focusing on the cheeks selects a region with abundant and relatively stable wrinkles.

[0102] Step 103: Calculate the vector similarity between the standard facial features and the real-time facial features; Quantify the similarity between the currently captured face and a registered face in the database. This step is crucial for converting image features into comparable numerical values.

[0103] Specifically, step 103 includes steps 1031 to 1035: Step 1031: Calculate the first similarity between the first standard facial features and the facial feature distribution features; Step 1032: Calculate the second similarity between the second standard face features and the real-time wrinkle texture features; Step 1033: If both the first similarity and the second similarity are greater than or equal to a preset similarity, then the minimum value between the first similarity and the second similarity is taken as the vector similarity. If both the first similarity and the second similarity are greater than or equal to a preset similarity (a set threshold, such as 0.87), then the minimum value between the first similarity and the second similarity is selected as the final vector similarity.

[0104] The purpose of this is to ensure that the features of both aspects reach a certain similarity standard, and that the final result is not obscured by a particularly high similarity, thus reflecting a more comprehensive similarity situation.

[0105] Step 1034: If the first similarity or the second similarity is less than the preset similarity, then the first similarity and the second similarity are weighted and summed to obtain the vector similarity; If either the first similarity or the second similarity is less than the preset similarity, then the two similarities are weighted and summed to obtain the final vector similarity.

[0106] The weighted summation method can adjust the influence ratio of the two similarities in the final result by setting different weights (e.g., 0.431 and 0.569), ensuring that even if one similarity is low, a reasonable similarity result can be obtained through comprehensive calculation.

[0107] Step 1035: If both the first similarity and the second similarity are less than the preset similarity, then the vector similarity is 0.

[0108] If both the first and second similarities are lower than the preset similarity, meaning that neither similarity meets the requirements, then the vector similarity is set to 0.

[0109] This indicates that neither the facial feature distribution characteristics nor the wrinkle texture characteristics achieved sufficient similarity with the standard features, and were therefore judged as dissimilar.

[0110] In the embodiments corresponding to steps 1031 to 1035, different similarity calculation and comparison methods are used to ensure the accuracy and rationality of similarity calculation, thereby providing reliable technical support for user management in the smart elderly care system. This method not only considers the comprehensive comparison of multiple features, but also ensures the robustness and accuracy of the system by setting a preset similarity threshold.

[0111] Step 104: When the vector similarity is greater than a preset value, the activity area corresponding to the camera device is used as user management information; wherein, the user management information is used to record the user's daily activity type.

[0112] If the calculated similarity is greater than a preset value (such as 0.8), it is determined that the currently captured face image is successfully matched with the user identity represented by the pre-stored standard feature.

[0113] Once identification is successful, the system records the user's identity information, confirming who the user is (corresponding to the standard features of a successful match). The system knows the specific location where the camera is installed (e.g., bedroom in room 101, or the east area of ​​the activity room on the second floor). Therefore, when a user is identified, the system also knows the user's current physical location (area). The confirmed user identity and their current location (area) are linked together as a user management record. The core purpose of this continuously generated user management information (user X appears in area Z at time Y) is to analyze the frequency, duration, and movement trajectory of users appearing in different functional areas (e.g., bedroom-sleep, dining room-dining, activity room-recreation and rehabilitation, bathroom-toilet) at different times. The system can automatically infer and record the user's main daily activity types (sleep, dining, socializing, rehabilitation training, toileting, etc.). These activity records are an important data foundation for health monitoring (e.g., sleep abnormalities, insufficient activity), safety alerts (e.g., prolonged stay in the bathroom), service optimization (e.g., adjusting meal delivery time based on habits), and personalized care.

[0114] In the embodiments corresponding to steps 101 to 104, real-time features of the face image to be identified are comprehensively extracted by combining facial feature distribution features and real-time wrinkle texture features. Compared with traditional methods that rely solely on static facial features, this invention can more accurately capture subtle facial features that change over time in the elderly, thereby improving the accuracy of face recognition, especially when facial features of the elderly change significantly. By calculating the vector similarity between standard facial features and real-time facial features, and making judgments based on a preset similarity threshold, this invention ensures the reliability of user identification. Even under different lighting conditions and changes in facial expressions, the system can still perform face recognition stably and accurately, ensuring the reliability and effectiveness of user management information. The activity area corresponding to the camera device is used as user management information to record the user's daily activity types in real time. This real-time user activity management function not only improves the timeliness of data, but also provides accurate user behavior data support for other service modules in the smart elderly care system, which helps to implement personalized service and care plans.

[0115] like Figure 2 This invention provides a user management device based on smart elderly care; please refer to [link / reference]. Figure 2 , Figure 2 The diagram shows a user management device based on smart elderly care provided by the present invention, such as... Figure 2 The user management device based on smart elderly care shown includes: The acquisition unit 21 is used to acquire images of the face to be identified through a camera device and obtain multiple pre-stored standard facial features; Extraction unit 22 is used to extract real-time facial features from the face image to be identified; the real-time facial features include facial feature distribution features and real-time wrinkle texture features; The calculation unit 23 is used to calculate the vector similarity between the standard face features and the real-time face features; The judgment unit 24 is used to take the activity area corresponding to the camera device as user management information when the vector similarity is greater than a preset value; wherein, the user management information is used to record the user's daily activity type.

[0116] This invention provides a user management device based on smart elderly care. By combining facial feature distribution characteristics and real-time wrinkle texture features, it comprehensively extracts real-time features from the face image to be identified. Compared to traditional methods that rely solely on static facial features, this invention can more accurately capture subtle facial features that change over time in the elderly, thereby improving the accuracy of face recognition, especially when facial features change significantly. By calculating the vector similarity between standard facial features and real-time facial features, and making judgments based on a preset similarity threshold, this invention ensures the reliability of user identification. Even under different lighting conditions and changes in facial expressions, the system can still perform face recognition stably and accurately, ensuring the reliability and effectiveness of user management information. The activity area corresponding to the camera device is used as user management information to record the user's daily activity types in real time. This real-time user activity management function not only improves the timeliness of data but also provides accurate user behavior data support for other service modules in the smart elderly care system, facilitating the implementation of personalized services and care plans.

[0117] Figure 3 This is a schematic diagram of a terminal device provided in an embodiment of the present invention. Figure 3 As shown, a terminal device 3 in this embodiment includes: a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30, such as a user management program based on smart elderly care. When the processor 30 executes the computer program 32, it implements the steps described in the above embodiments of a user management method based on smart elderly care, for example... Figure 1 Steps 101 to 104 are shown. Alternatively, when the processor 30 executes the computer program 32, it implements the functions of each unit in the above-described device embodiments, for example... Figure 2 The function of the unit shown.

[0118] For example, the computer program 32 can be divided into one or more units, which are stored in the memory 31 and executed by the processor 30 to complete the present invention. The one or more units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program 32 in the terminal device 3. For example, the specific functions of each unit of the computer program 32 can be divided as follows: The acquisition unit is used to acquire images of the face to be identified through a camera device and obtain multiple pre-stored standard facial features; An extraction unit is used to extract real-time facial features from the face image to be identified; the real-time facial features include facial feature distribution features and real-time wrinkle texture features; A calculation unit is used to calculate the vector similarity between the standard facial features and the real-time facial features; The judgment unit is used to determine the activity area corresponding to the camera device as user management information when the vector similarity is greater than a preset value; wherein the user management information is used to record the user's daily activity type.

[0119] The terminal device includes, but is not limited to, a processor 30 and a memory 31. Those skilled in the art will understand that... Figure 3 This is merely an example of a terminal device 3 and does not constitute a limitation on a terminal device 3. It may include more or fewer components than shown, or combine certain components, or different components. For example, the terminal device may also include input / output devices, network access devices, buses, etc.

[0120] The processor 30 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0121] The memory 31 can be an internal storage unit of the terminal device 3, such as a hard disk or memory of the terminal device 3. The memory 31 can also be an external storage device of the terminal device 3, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the terminal device 3. Furthermore, the memory 31 can include both internal and external storage units of the terminal device 3. The memory 31 is used to store the computer program and other programs and data required by the roaming control device. The memory 31 can also be used to temporarily store data that has been output or will be output.

[0122] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0123] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of the present invention. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0124] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this invention. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0125] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0126] This invention provides a computer program product that, when run on a mobile terminal, enables the mobile terminal to implement the steps described in the above-described method embodiments.

[0127] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying the computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0128] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0129] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0130] In the embodiments provided by this invention, it should be understood that the disclosed apparatus / network devices and methods can be implemented in other ways. For example, the apparatus / network device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0131] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units.

[0132] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0133] It should also be understood that the term “and / or” as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0134] As used in this specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if [the described condition or event] is detected" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once [the described condition or event] is detected," or "in response to detection of [the described condition or event]."

[0135] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0136] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0137] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A user management method based on smart elderly care, characterized in that, The user management method based on smart elderly care includes: The camera device captures images of the faces to be identified and obtains multiple pre-stored standard facial features; Extract real-time facial features from the face image to be identified; the real-time facial features include facial feature distribution features and real-time wrinkle texture features; Calculate the vector similarity between the standard facial features and the real-time facial features; When the vector similarity is greater than a preset value, the activity area corresponding to the camera device is used as user management information; wherein, the user management information is used to record the user's daily activity type.

2. The user management method based on smart elderly care as described in claim 1, characterized in that, The step of extracting real-time facial features from the image of the face to be identified includes: The facial contours in the image of the person to be identified are extracted based on the Haar cascade classifier; The left corner of the eye, the right corner of the eye, and the tip of the nose in the image of the face to be identified are extracted using a facial landmark detector. Based on the distribution characteristics of the facial contour, the position of the left corner of the eye, the position of the right corner of the eye, and the position of the tip of the nose, the facial feature distribution characteristics are constructed. Extract a local facial region located between the corner of the eye and the chin from the face region, and remove the preset nose region from the local facial region to obtain the cheek region; wherein, the corner of the eye refers to the straight line formed by the left corner of the eye and the right corner of the eye; Extract real-time wrinkle texture features from the cheek region; The facial feature distribution features and the real-time wrinkle texture features are used as the real-time face features.

3. The user management method based on smart elderly care as described in claim 2, characterized in that, The step of constructing the facial feature distribution features based on the distribution characteristics between the facial contour, the position of the left eye corner, the position of the right eye corner, and the position of the nose tip includes: Calculate the first distance between the tip of the nose and the left corner of the eye, and calculate the second distance between the tip of the nose and the right corner of the eye; Calculate the average distance between the first distance and the second distance; A circular region is constructed with the nose tip position as the center and the average distance as the radius; Extract the intersection point between the circular region and the facial contour; Calculate the third distance between the intersection point and the positions of the left corner of the eye, the right corner of the eye, and the tip of the nose, respectively; Calculate the fourth distance between the positions of the left and right corners of the eyes; The first distance, the second distance, multiple third distances, and the fourth distance are normalized and combined in a fixed order to form the facial feature distribution.

4. The user management method based on smart elderly care as described in claim 2, characterized in that, The step of extracting real-time wrinkle texture features of the cheek region includes: The cheek area is processed to obtain a binarized image; Extract the texture region corresponding to the maximum value in the binarized image; For each texture region, extract the first pixel located at the leftmost edge of the texture region, extract the second pixel located at the rightmost edge of the texture region, and extract the third pixel located at the center of the texture region; The first pixel, the second pixel, and the third pixel are designated as marker pixels. For each marker pixel, extract the pixel neighborhood of a preset size around that marker pixel; Extract target texture units from the pixel neighborhood; wherein, the target texture unit is a set of pixels with consistent features; Real-time wrinkle texture features are extracted based on the target texture unit.

5. The user management method based on smart elderly care as described in claim 4, characterized in that, The step of extracting the target texture unit in the pixel region includes: Based on pixel neighborhood, directional response intensity and dominant polarity are extracted in multiple preset directions; wherein, the multiple preset directions include 0°, 45°, 90° and 135°; the directional response intensity is used to quantify the significance of structural changes in the pixel neighborhood in the preset directions; the dominant polarity is used to quantify the brightness transition trend in the pixel neighborhood in the preset directions. The pixel corresponding to the maximum directional response intensity is used as the seed point, and the neighboring pixels around the seed point are obtained. When the first adjacent pixel of the seed point meets the preset conditions, the seed point and the first adjacent pixel are merged into a texture unit; wherein, the preset conditions include that the dominant polarity of the two is the same and the difference in directional response intensity between the two is less than a threshold. When the second adjacent pixel of a texture unit meets the preset conditions, the texture unit and the second adjacent pixel are merged into a new texture unit; The step of merging the texture unit and the second adjacent pixel into a new texture unit when the second adjacent pixel of the texture unit meets the preset conditions is executed iteratively to obtain the target texture unit.

6. The user management method based on smart elderly care as described in claim 4, characterized in that, The step of extracting real-time wrinkle texture features based on the target texture unit includes: Calculate the first perimeter of the target texture unit; Construct a circle with the same area as the target texture unit, and extract the second circumference of the circle; Subtracting the first perimeter from the second perimeter and dividing by the second perimeter yields the boundary irregularity; wherein the boundary irregularity is used to measure the degree of boundary irregularity. The statistical analysis identifies the enclosed hole regions within the target texture unit; where a hole region refers to an independent region enclosed by the target texture unit. Divide the number of pixels in all hole regions by the total number of pixels in the target texture unit to obtain the percentage of hole regions; In the binarized image, for each current pixel in the target texture unit, obtain the eight adjacent pixels of that current pixel; When all eight pixels are pixels in the target texture unit, the current pixel is taken as a connected component; Divide the number of connected components by the total number of pixels in the target texture unit to obtain the proportion of connected components; Extract the center point of the texture region, calculate the first angle between the center point and the left corner of the eye, and calculate the second angle between the center point and the right corner of the eye; The boundary irregularity, the proportion of the hole region, the proportion of the connected components, the first included angle, and the second included angle are used as real-time wrinkle texture features.

7. The user management method based on smart elderly care as described in claim 1, characterized in that, The standard facial features include a first standard facial feature and a second standard facial feature; The step of calculating the vector similarity between the standard facial features and the real-time facial features includes: Calculate the first similarity between the first standard facial features and the facial feature distribution features; Calculate the second similarity between the second standard face features and the real-time wrinkle texture features; If both the first similarity and the second similarity are greater than or equal to a preset similarity, then the minimum value between the first similarity and the second similarity is taken as the vector similarity. If the first similarity or the second similarity is less than the preset similarity, then the first similarity and the second similarity are weighted and summed to obtain the vector similarity. If both the first similarity and the second similarity are less than the preset similarity, then the vector similarity is 0.

8. A user management device based on smart elderly care, characterized in that, The smart elderly care-based user management device includes: The acquisition unit is used to acquire images of the face to be identified through a camera device and obtain multiple pre-stored standard facial features; An extraction unit is used to extract real-time facial features from the face image to be identified; the real-time facial features include facial feature distribution features and real-time wrinkle texture features; A calculation unit is used to calculate the vector similarity between the standard facial features and the real-time facial features; The judgment unit is used to determine the activity area corresponding to the camera device as user management information when the vector similarity is greater than a preset value; wherein the user management information is used to record the user's daily activity type.

9. A terminal device, characterized in that, The terminal device includes: a memory, a processor, and a smart elderly care-based user management program stored in the memory and executable on the processor, wherein the smart elderly care-based user management program is configured to implement the steps in the smart elderly care-based user management method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps in the user management method based on smart elderly care as described in any one of claims 1 to 7.

Images