Mask determination method, apparatus, device, and readable storage medium

Abstract

This application provides a method, apparatus, device, and readable storage medium for determining a face mask. The method includes: recognizing a user's facial image to obtain facial features in the facial image; each facial feature represents the distance between two feature points in the facial image; when there is a reference object in the facial image, determining a target ratio between the projection size of the reference object in the facial image and the facial features, and determining the target ratio as a first type; the first type represents the size relationship between the projection size and the facial features; when there is no reference object in the facial image, determining a target ratio between two facial features, and determining the target ratio as a second type; the second type represents the size relationship between the two facial features; determining a mapping relationship corresponding to the ratio type of the target ratio, and determining the target face mask specification corresponding to the target ratio in the mapping relationship. This application can improve the efficiency of face mask determination.

Description

Technical Field

[0001] This disclosure relates to the field of text management technology, and in particular to a method, apparatus, electronic device, and computer-readable storage medium for determining a face mask. Background Technology

[0002] Face masks fit snugly to the face, and different masks serve different purposes, such as medical, protective, or decorative functions. The fit of the mask directly affects its functionality. For example, in modern medical fields, the fit between the mask of a ventilation therapy device and the user can affect the treatment outcome.

[0003] In related technologies, the eyes in a facial image are used as a fixed reference. Based on the measurement value of the eyes in the facial image and the preset eye size, the image scaling degree is determined. Based on the image scaling degree and the measurement values ​​of other parts in the facial image, the size of other parts is determined, and the corresponding mask is determined based on the size of other parts.

[0004] Therefore, in related technologies, the selection of a face mask involves a cumbersome calculation process involving eye measurements, eye size, image scaling, measurements of other parts of the body, and sizes of other parts of the body, which is inefficient. Summary of the Invention

[0005] In view of the above problems, embodiments of the present disclosure are proposed to provide a mask determination method, apparatus, electronic device and computer-readable storage medium that overcomes or at least partially solves the above problems.

[0006] In a first aspect, embodiments of this disclosure provide a method for determining a face mask, including:

[0007] Recognize the user's facial image to obtain facial features in the facial image; each facial feature represents the distance between two feature points in the facial image;

[0008] In the case of a reference object in the facial image, a target ratio between the projection size of the reference object in the facial image and the facial features is determined, and the target ratio is determined to be of a first type; the first type represents the size relationship between the projection size and the facial features.

[0009] In the absence of a reference object in the facial image, a target ratio between two facial features is determined, and the target ratio is identified as a second type; the second type represents the size relationship between the two facial features.

[0010] Determine the mapping relationship corresponding to the proportion type of the target proportion, and determine the target mask size corresponding to the target proportion in the mapping relationship; different proportion types correspond to different mapping relationships; the mapping relationship is the correspondence between proportion and mask size.

[0011] Secondly, embodiments of this disclosure disclose a mask determining device, comprising:

[0012] An image recognition module is used to recognize a user's facial image to obtain facial features in the facial image; each facial feature represents the distance between two feature points in the facial image;

[0013] The reference ratio module is used to determine, when there is a reference object in the facial image, the target ratio between the projection size of the reference object in the facial image and the facial features, and to determine that the target ratio is a first type; the first type represents the size relationship between the projection size and the facial features.

[0014] A facial proportion module is used to determine a target proportion between two facial features when there is no reference object in the facial image, and to determine the target proportion as a second type; the second type represents the size relationship between the two facial features.

[0015] The mask determination module is used to determine a mapping relationship corresponding to the proportion type of the target proportion, and to determine the target mask specification corresponding to the target proportion in the mapping relationship; different proportion types correspond to different mapping relationships; the mapping relationship is the correspondence between proportion and mask specification.

[0016] Thirdly, this disclosure also discloses an electronic device, including a processor and a memory, wherein the memory stores a program or instructions executable on the processor, and the program or instructions, when executed by the processor, implement the steps of the mask determination method as described in the first aspect.

[0017] Fourthly, embodiments of this disclosure also disclose a computer-readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the mask determination method as described in the first aspect.

[0018] In this embodiment, facial features are obtained by recognizing a user's facial image. Each facial feature represents the distance between two feature points in the facial image. When there is a reference object in the facial image, the target ratio between the projection size of the reference object in the facial image and the facial features is determined, and the target ratio is determined to be of type 1. When there is no reference object in the facial image, the target ratio between two facial features is determined, and the target ratio is determined to be of type 2. The corresponding target ratio can be obtained under different conditions with or without a reference object, and the target mask specification indicated by the target ratio is determined in the mapping relationship corresponding to the ratio type. The corresponding target mask specification can be determined directly based on the corresponding target ratio, avoiding the tedious calculation process involving the measurement values, sizes, image scaling and other contents of different facial features, thus improving the efficiency of mask determination. At the same time, since the mask is determined in real time based on the actual ratio between features in the facial image, rather than using the fixed preset size of a specific facial feature as a reference, the mask determination process also has high flexibility, thereby improving the accuracy of mask determination. Attached Figure Description

[0019] Figure 1 This is a step diagram of a mask determination method provided in an embodiment of this disclosure;

[0020] Figure 2 This is a schematic diagram of facial features provided in an embodiment of this disclosure;

[0021] Figure 3 This is a schematic diagram of a facial image including a reference object provided in an embodiment of this disclosure;

[0022] Figure 4 This is a step diagram of another mask determination method provided in this embodiment of the disclosure;

[0023] Figure 5 This is a schematic diagram of the feature orientation provided in the embodiments of this disclosure;

[0024] Figure 6 This is a schematic diagram showing the tilt of the reference object provided in the embodiments of this disclosure;

[0025] Figure 7 This is another schematic diagram showing the tilt of the reference object provided in the embodiments of this disclosure;

[0026] Figure 8 This is a block diagram of a mask determining device provided in an embodiment of this disclosure;

[0027] Figure 9 This is a block diagram of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0028] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0029] In modern medicine, the treatment of respiratory diseases increasingly relies on ventilation therapy equipment, especially the fit of face masks, which directly affects treatment outcomes. Mask fit not only impacts patient comfort but also the effective delivery of airflow and therapeutic efficacy. Traditional ventilation mask designs often employ standardized sizes and shapes, but each patient's facial structure varies significantly, including mouth thickness, eye spacing, and nose height. Fixed, universal designs can easily lead to a mismatch between the mask and facial contours, causing airflow leakage, reducing treatment efficiency, and increasing patient discomfort.

[0030] Figure 1 The diagram illustrates the steps of a mask determination method provided in an embodiment of this disclosure. The method includes:

[0031] Step 101: Recognize the user's facial image to obtain facial features in the facial image; each facial feature represents the distance between two feature points in the facial image.

[0032] In this embodiment of the disclosure, the mask determination method step can be executed by a terminal device. The terminal device can be configured with an image acquisition device, such as a camera or scanner. The user can operate the terminal device to capture their own facial image through the image acquisition device, such as taking a picture of the head or upper body to obtain a facial image.

[0033] During the process of recognizing a user's facial image, some preprocessing can be performed, such as grayscale processing, which can reduce the amount of computation while retaining sufficient information, or noise reduction processing can be used to reduce noise in the image. There are no specific restrictions on the preprocessing operations before recognizing the image.

[0034] When recognizing facial images, key facial points, or feature points, such as the corners of the eyes, the tip of the nose, the nostrils, and the corners of the mouth, can be identified. Feature point recognition can be achieved through methods such as Active Shape Model (ASM), Active Appearance Model (AAM), Cascade Regression (CR) algorithms, or Convolutional Neural Networks (CNNs). For example, when using ASM, a shape model can be first built. By statistically analyzing facial feature points in training samples, the average shape and principal components of shape changes can be obtained. During the recognition process, an iterative search is used to match the model shape with the facial shape in the image, thereby locating the feature points.

[0035] Facial features can be defined based on the distances between different feature points. For example, the facial feature "mouth thickness" could represent the distance between the midpoint of the upper edge of the mouth (feature point) and the midpoint of the lower edge of the mouth (feature point), while the facial feature "nose width" could represent the distance between the left and right nasal alar endpoints (feature points). Facial features can also be the distance between the center of the pupil and the tip of the nose; there are no specific restrictions on the distances corresponding to different facial features.

[0036] Figure 2 This is a schematic diagram of facial features provided in an embodiment of this disclosure; Figure 2 State A represents the user's closed mouth. State A includes facial features D1 (mouth thickness) 201, D2 (nose width) 202, and D3 (height from corner of mouth to eyebrow) 203. In D3, the eyebrow position can be defined as the point near the center of the brow. State B represents the user's open mouth. When the mouth is open, key points include the midpoint between the upper and lower edges of the upper lip, and the midpoint between the upper and lower edges of the lower lip. Mouth thickness is the distance between the midpoints of the upper and lower edges of the upper lip, plus the distance between the midpoints of the upper and lower edges of the lower lip.

[0037] Understandably, each facial feature represents the distance between two feature points, which can be horizontal or vertical. For different facial features to be distinct, at least one feature point must be different. For example, facial feature 1 might represent the distance between feature points A and B, facial feature 2 might represent the distance between feature points AC, and facial feature 3 might represent the distance between feature points CF.

[0038] Step 102: If there is a reference object in the facial image, determine the target ratio between the projection size of the reference object in the facial image and the facial features, and determine the target ratio as a first type; the first type represents the size relationship between the projection size and the facial features.

[0039] In this embodiment of the disclosure, a reference object may be present in the facial image, and the projection size of the reference object in the facial image can be determined. The projection size represents the distance between the two endpoints of the reference object. For example, if the reference object is a rectangular card, the projection size can be the distance between any two corner points, such as the distance between two adjacent corner points, i.e., the side length.

[0040] Figure 3 This is a schematic diagram of a facial image including a reference object provided in an embodiment of this disclosure. When the user 301 is horizontally facing the image acquisition device at a standard angle, the reference object 302 can be pressed against the chin of the user 301's face when acquiring the facial image. Since the reference object 302 is close to the face, it can be understood that the reference object 302 and the user 301's face are on the same plane.

[0041] If both the reference object and the facial feature exist in the same facial image, the projected size of the reference object and the distance represented by the facial feature can be compared to determine the target ratio between the projected size and the facial feature. This type of target ratio can be defined as the first type, which represents the size relationship between the projected size and the facial feature.

[0042] Step 103: In the absence of a reference object in the facial image, determine the target ratio between two facial features and determine the target ratio as a second type; the second type represents the size relationship between two facial features.

[0043] In this embodiment of the disclosure, multiple facial features can be obtained from the facial image, and each facial feature represents the distance between two feature points. When there is no reference object in the facial image, two facial features can be selected from the multiple facial features, and the ratio between these distances, based on the corresponding distances in each facial feature, is the target ratio.

[0044] The target ratio can represent the size relationship between two facial features in the target ratio. For example, if the distance corresponding to facial feature 1 is 20 and the distance corresponding to facial feature 2 is 10, then the target ratio is 2, which can mean that the size of facial feature 1 is twice that of facial feature 2.

[0045] Step 104: Determine the mapping relationship corresponding to the ratio type of the target ratio, and determine the target mask specification corresponding to the target ratio in the mapping relationship; different ratio types correspond to different mapping relationships; the mapping relationship is the correspondence between ratio and mask specification.

[0046] In this embodiment of the disclosure, since the target ratio represents the size relationship between the reference object and the facial features, or between two facial features, the target ratio actually represents the user's facial structure, and the mask size suitable for the user's facial structure can be selected based on the target ratio.

[0047] A mapping relationship between mask specifications and proportions can be pre-established, where different mask specifications represent different mask sizes, such as small, medium, and large. The target mask specification can then be determined based on the target proportion, thus completing the mask selection.

[0048] In particular, since the facial proportions between the reference object and facial features, and the target proportions between two facial features are based on measurements of the facial structure from different angles, different mapping relationships are required, and mapping relationships with the same data content cannot be used.

[0049] Once the target mask specifications are determined, they can be displayed on the device's display interface for recommendation to the user.

[0050] In summary, in this embodiment of the present disclosure, facial features are obtained by recognizing a user's facial image. Each facial feature represents the distance between two feature points in the facial image. When there is a reference object in the facial image, the target ratio between the projection size of the reference object in the facial image and the facial features is determined, and the target ratio is determined to be of the first type. When there is no reference object in the facial image, the target ratio between two facial features is determined, and the target ratio is determined to be of the second type. The corresponding target ratio can be obtained under different conditions with or without a reference object, and the target mask specification indicated by the target ratio is determined in the mapping relationship corresponding to the ratio type. The corresponding target mask specification can be determined directly based on the corresponding target ratio, avoiding the tedious calculation process involving the measurement values, sizes, image scaling and other contents of different facial features, thus improving the efficiency of mask determination. At the same time, since the mask is determined in real time based on the actual ratio between features in the facial image, rather than using the fixed preset size of a specific facial feature as a reference, the mask determination process also has high flexibility, thereby improving the accuracy of mask determination.

[0051] refer to Figure 4 The document illustrates a flowchart of a mask determination method provided in an embodiment of this disclosure, the method comprising:

[0052] Step 401: Recognize the user's facial image to obtain facial features in the facial image; each facial feature represents the distance between two feature points in the facial image;

[0053] Step 402: If there is a reference object in the facial image, determine the target ratio between the projection size of the reference object in the facial image and the facial features, and determine the target ratio as a first type; the first type represents the size relationship between the projection size and the facial features.

[0054] Step 403: In the absence of a reference object in the facial image, determine the target ratio between two facial features and determine the target ratio as a second type; the second type represents the size relationship between two facial features.

[0055] Step 404: Determine the mapping relationship corresponding to the ratio type of the target ratio, and determine the target mask specification corresponding to the target ratio in the mapping relationship; different ratio types correspond to different mapping relationships; the mapping relationship is the correspondence between ratio and mask specification.

[0056] Steps 401-404 above can be referred to the above. Figure 1 The details of the embodiments will not be repeated here.

[0057] In this embodiment of the disclosure, the face mask can be classified according to the area of ​​the face it covers. The coverage type may include, for example, a full-face mask, a nose mask, and a nose pillow mask. A full-face mask can cover the entire face, a nose mask can cover the entire nose, and a nose pillow mask can only cover the nostrils.

[0058] Each type of face mask can correspond to different mask sizes. For example, full-face masks come in small, medium, and large sizes, as do nose masks. However, for the same size, different coverage types correspond to different proportions. For instance, with a proportion of 0.7, a suitable full-face mask size is medium, and a suitable nose mask size is small. Therefore, within a mapping relationship between proportion and mask size, different coverage types can have different sub-mapping relationships. Although each sub-mapping relationship is a correspondence between mask size and proportion, the specific details of the sub-mapping relationship can differ.

[0059] Users can trigger actions on the device's display interface to determine the required coverage type, such as a full-face mask. This allows the determination of the target sub-mapping relationship corresponding to the user-indicated target coverage type, and within that sub-mapping relationship, the target mask specification corresponding to the target proportion.

[0060] Understandably, users can also choose not to specify a particular target coverage type. In this case, the device can determine the target mask specifications for each coverage type based on the target ratio and display them on the device's screen for recommendation to the user.

[0061] By implementing embodiments of this disclosure, by obtaining the target coverage type indicated by the user, and determining the target mask specification corresponding to the target ratio in the target mapping relationship between the mask specification and ratio corresponding to the target coverage type, the corresponding target mask specification can be determined according to the specific coverage type required by the user, thereby improving the accuracy of mask determination.

[0062] Optionally, step 404, which involves determining the mapping relationship between the proportion type and the target proportion, and determining the target mask specification corresponding to the target proportion within the mapping relationship, includes:

[0063] Sub-step 4041: If there are at least two target ratios, determine a set of mapping relationships corresponding to the ratio types of the target ratios; the ratio types of at least two target ratios are either both of the first type or both of the second type; the set of mapping relationships includes at least two mapping relationships.

[0064] Sub-step 4042: For different target ratios, determine the target mask specifications in different mapping relationships to obtain at least two target mask specifications.

[0065] In this embodiment of the disclosure, there may be at least two target proportions, and the proportion types of the target proportions are either all of the first type or all of the second type. This is because there are only two scenarios: with a reference object or without a reference object. If there is a reference object, then even if there are multiple target proportions, they are all of the first type; if there is no reference object, then even if there are multiple target proportions, they are all of the second type.

[0066] Each target ratio corresponds to a different mapping relationship. This is because different target ratios actually correspond to different facial features. For example, the first target ratio of the first type could be about the projection size and facial feature A, and the second target ratio of the first type could be about the projection size and facial feature B. Since the projection size is the same, different facial features are needed to have different target ratios of the first type. The target ratio of the second type is similar. If the two target ratios of the second type correspond to two completely identical facial features, then they are actually the same target ratio.

[0067] Therefore, since different target proportions are determined by different facial features, different types of facial proportions are also measured from different angles of facial structure, thus requiring different mapping relationships.

[0068] The number of mapping relationships in the mapping relationship set is consistent with the number of target proportions, which allows the corresponding target mask specifications to be determined in different mapping relationships for each target proportion.

[0069] Optionally, there are two target proportions, both of which are of the first type; different mapping relationships in the mapping relationship set correspond to different facial features;

[0070] The step of determining the target mask specifications in different mapping relationships for different target proportions to obtain two target mask specifications includes:

[0071] Determine the first facial feature corresponding to the first target ratio and the first mapping relationship corresponding to the first facial feature, and determine the first target mask specification corresponding to the first target ratio in the first mapping relationship;

[0072] Determine the second facial feature corresponding to the second target ratio, and the second mapping relationship corresponding to the second facial feature, and in the second mapping relationship, determine the second target mask specification corresponding to the second target ratio;

[0073] The feature directions of the first facial feature and the second facial feature are perpendicular to each other; the feature direction is determined by the line connecting the two feature points.

[0074] In this embodiment, there can be two target proportions, both of which are of the first type, belonging to a scenario with a reference object. In a scenario with a reference object, the target proportion is obtained based on the projection size and a facial feature. The first target proportion of the two target proportions corresponds to a first facial feature, and different facial features correspond to different mapping relationships, which will not be elaborated further.

[0075] Then, within the first mapping relationship corresponding to the first facial feature, the first target mask specification corresponding to the first target ratio can be determined. Similarly, the second facial feature corresponding to the second target ratio and the second mapping relationship corresponding to the second facial feature are determined, and within the second mapping relationship, the second target mask specification corresponding to the second target ratio is determined.

[0076] Since facial scanning acquires a two-dimensional image, it cannot determine the true size. Therefore, selecting facial features in different dimensions, i.e., different feature directions, helps improve measurement accuracy. The feature directions of the first and second facial features are perpendicular to each other. The feature direction can be defined as the line connecting the two feature points corresponding to each facial feature. For example, the feature direction of the facial feature of mouth thickness is vertical, while the feature direction of the facial feature of nose width is horizontal.

[0077] Generally, facial features can be horizontal or vertical, meaning the horizontal and vertical directions are perpendicular. However, other directions are also possible. For example, using the distance between the outer corner of the eye and the tip of the nose as a facial feature results in a tilted direction. Similarly, several facial features with different directions can be obtained. If a tilted direction is used, the target proportion can be multiple, such as five or six. Generally, the maximum difference between two feature directions (i.e., when they are perpendicular) can be used to determine the target mask specifications.

[0078] Figure 5 This is a schematic diagram of the feature orientation provided in the embodiments of this disclosure; Figure 5 It can include two feature directions, or two dimensions: a first dimension in the horizontal direction and a second dimension in the vertical direction.

[0079] The first target mask specification determined based on the first target ratio and the first mapping relationship, and the second target mask specification determined based on the second target ratio and the second mapping relationship, can be understood as results obtained from different measurement methods. Each result can reflect the user's facial structure from a specific angle. Therefore, if the results of the first target mask specification and the second target mask specification are consistent, a mask determination result with higher accuracy can be determined and recommended to the user. If the two results are inconsistent, both results can be recommended to the user for selection.

[0080] By implementing the embodiments of this disclosure, facial mask specifications can be determined in different mapping relationships by using two feature directions of two target proportions that are perpendicular to each other for facial features. Target mask specifications can be determined separately from different feature directions, thereby improving the comprehensiveness and accuracy of mask determination.

[0081] Optionally, there are two target proportions, both of which are of the second type; different mapping relationships in the mapping relationship set correspond to different facial features;

[0082] The step of determining the target mask specifications in different mapping relationships for different target proportions to obtain two target mask specifications includes:

[0083] Determine the first facial feature and the second facial feature corresponding to the first target ratio, as well as the first mapping relationship corresponding to the first facial feature and the second facial feature, and determine the first target mask specification corresponding to the first target ratio in the first mapping relationship;

[0084] Determine the third and fourth facial features corresponding to the second target ratio, and the second mapping relationship corresponding to the third and fourth facial features, and in the second mapping relationship, determine the second target mask specification corresponding to the second target ratio;

[0085] The first facial feature and the third facial feature have different feature directions; the second facial feature and the fourth facial feature have different feature directions.

[0086] In this embodiment of the disclosure, in a scenario without a reference object, there are two target proportions, both of which are of the second type. The first target proportion corresponds to a first facial feature and a second facial feature, and the second target proportion corresponds to a third facial feature and a fourth facial feature. At least one of the facial features corresponding to the different target proportions is different.

[0087] For example, the first target ratio corresponds to facial feature 1 and facial feature 2, the second target ratio corresponds to facial feature 1 and facial feature 2 (with one difference), and the second target ratio can also correspond to facial feature 2 and facial feature 3 (with two differences). See also... Figure 2 If the first target ratio corresponds to facial features D1 and D3, then the second target ratio corresponds to facial features D2 and D3.

[0088] The feature directions of the first and second facial features are perpendicular to each other, as are the feature directions of the third and fourth facial features. For example, the feature direction of the first facial feature is horizontal, the feature direction of the second facial feature is vertical, the feature direction of the third facial feature is vertical, and the feature direction of the fourth facial feature is horizontal. This allows the first target ratio to represent the facial features in the horizontal direction divided by the facial features in the vertical direction, and the second target ratio to represent the facial features in the vertical direction divided by the facial features in the horizontal direction. This enables the facial structure to be determined from two opposite directions.

[0089] Since the facial features corresponding to the two target proportions are at least partially different, and the target proportions corresponding to the two different facial features represent the facial structure from different angles, different target proportions need to correspond to their respective mapping relationships. The mapping relationship is still the correspondence between mask specifications and proportions.

[0090] For example, if the facial features included in the first target proportion are nose width and mouth thickness, corresponding to mapping relationship A, and the facial features included in the second target proportion are eye width and mouth thickness, then another mapping relationship B needs to be applied, because the facial structure expressed by the ratio between nose width and mouth thickness is different from the facial structure expressed by the ratio between eye width and mouth thickness.

[0091] Based on the first facial features and the second facial features, a first target mask specification corresponding to the first target proportion can be determined in a first mapping relationship. Similarly, based on the second target proportion and the second target proportion, a second target mask specification can be determined in a second mapping relationship.

[0092] Understandably, for two scenarios with or without reference points, if two target scales exist, the system can simultaneously receive the target coverage type indicated by the user and determine the target mask specifications based on the two target scales respectively. See Tables 1 and 2 below:

[0093] Table 1

[0094] Face mask specifications Full face mask (D1 / D3) Nose mask (D1 / D3) Nose pillow mask (D1 / D3) Small size >0.23 >0.55 >0.57 Medium 0.2～0.23 0.46～0.55 0.4～0.57 Plus Size <0.2 0.4～0.46 <0.4

[0095] Table 2

[0096] Face mask specifications Full face mask (D3 / D2) Nose mask (D3 / D2) Nose pillow mask (D3 / D2) Small size >0.23 >0.53 >0.57 Medium 0.2～0.23 0.45～0.53 0.4～0.57 Plus Size <0.2 <0.45 <0.4

[0097] First, based on the user's instructions, determine the target coverage type in Tables 1 and 2 above, such as a full-face mask, a nose mask, or a nose pillow mask. Then, based on the two selected target proportions, if the facial features corresponding to the first target proportion are D1 and D3, the target mask specifications can be determined from the sub-mapping relationship (one column of the table) corresponding to the target coverage type in Table 1.

[0098] If the facial features corresponding to the second ratio are D2 and D3, then based on the second ratio, the target mask specifications can be determined in the sub-mapping relationship corresponding to the target coverage type in Table 2.

[0099] By implementing the embodiments of this disclosure, when there are two target ratios, for the first target ratio, a first target mask specification corresponding to the first target ratio is determined in a first mapping relationship; for the second target ratio, a second target mask specification corresponding to the second target ratio is determined in a second mapping relationship; and then, based on the first target mask specification and the second target mask specification, the target mask specification is determined. This allows for the comprehensive determination of the mask specification using different facial features based on the two target ratios, thereby improving the accuracy of mask determination.

[0100] Optionally, the distance represented by each facial feature is a horizontal distance or a vertical distance; the step of determining the target ratio between two facial features when there is no reference object in the facial image includes:

[0101] Among multiple facial features in the facial image, two target facial features with different feature directions are identified; the two feature directions include a horizontal direction corresponding to the horizontal distance and a vertical direction corresponding to the vertical distance.

[0102] In the two target facial features, a reference facial feature is determined based on a preset stability parameter for each facial feature; features other than the reference facial feature are contrast facial features.

[0103] The target ratio between the two target facial features is determined based on the ratio between the first distance corresponding to the reference facial feature and the second distance corresponding to the comparison facial feature.

[0104] In this embodiment of the disclosure, the distance represented by each facial feature is either a horizontal distance or a vertical distance. For example, the thickness of the mouth represents the vertical distance, and the width of the nose represents the horizontal distance.

[0105] Select two facial features from multiple facial features that have different feature directions. These could be horizontal distances corresponding to the horizontal direction and vertical distances corresponding to the vertical direction. For example, the thickness of the mouth represents the vertical distance, and the feature direction is vertical; the width of the nose represents the horizontal distance, and the feature direction is horizontal.

[0106] Based on preset stability parameters for each facial feature, a reference facial feature and a comparison facial feature can be determined. When determining the reference facial feature, different facial features can have different priorities, and these priorities can be positively correlated with the stability parameters. For example, since the difference in mouth thickness among different users is usually small, a larger preset stability parameter can be used. If mouth thickness is included in the target proportion, then mouth thickness can be used as the reference facial feature. Similarly, the difference in nose width among different users can also usually be small, and a larger preset stability parameter can also be used. Therefore, the preset stability parameters of different facial features can be comprehensively determined to identify the reference facial feature, thereby improving the accuracy of mask selection for different users.

[0107] The distance represented by facial features, i.e., the first distance, can be used as a reference value. Then, a second distance represented by facial features is measured and compared. Based on the ratio of the first distance to the second distance, the target proportion is determined. When the first distance is used as a reference value, the reference value can be set to 1. By measuring the second distance and obtaining its relative size to the reference value, the facial proportion can be obtained.

[0108] Optionally, the reference object is a rectangle; when there is a reference object in the facial image, the step of determining the target ratio between the projection size of the reference object in the facial image and the facial features includes:

[0109] In the case where there is a reference object in the facial image and the reference object is tilted, the first tilt angle of the reference object in the horizontal direction is determined based on the actual width of the reference object and the projected width in the facial image.

[0110] The second tilt angle of the reference object in the vertical direction is determined based on the actual length of the reference object and its projected length in the facial image;

[0111] Based on the first tilt angle and the second tilt angle, the projection size of the reference object in the facial image is corrected, and a target ratio between the projection size and the facial features is determined.

[0112] In this embodiment of the disclosure, when capturing facial images, since the user may hold a reference object, the reference object will inevitably have a certain tilt angle. When the reference object is tilted, the projected value left by it in the image acquisition device will have a certain difference from its actual size. The size of the reference object can be known, for example, the reference object is a rectangular bank card or ID card, etc., with a fixed size. For example, the size of a bank card is 85.60 mm × 53.98 mm, with an aspect ratio of 4:3.

[0113] The size of the reference object can be deduced from its display range in the facial image and the calibration parameters of the image acquisition device. If the deduced result differs from the known size of the reference object, then the reference object is tilted.

[0114] Taking a bank card as a reference, the angle between the bank card and the plane (the vertical plane parallel to the image acquisition device) can be decomposed into tilt angles in two directions.

[0115] Figure 6 This is a schematic diagram showing the tilt of a reference object provided in an embodiment of this disclosure; as shown Figure 6 As shown, Figure 6 The face of user 601 can be on a plane, and the reference object 602 is tilted, including: a first tilt angle of θ about the width direction, i.e. the horizontal direction, and a second tilt angle of φ about the length direction, i.e. the vertical direction.

[0116] Given the actual width W and height H of the bank card, the projected width W' and length H' of the bank card in the facial image, the first tilt angle θ (around the width direction, i.e., the horizontal direction), and the second tilt angle φ (around the length direction, i.e., the vertical direction), according to the projection formula:

[0117] W=W′cosθ

[0118] H=H′cosφ

[0119] Therefore, we can conclude that:

[0120]

[0121] The values ​​of θ and φ can be calculated using the above formulas. After calculating the card's tilt angle, perspective transformation can be used to correct the projection. A transformation matrix can be calculated based on the tilt angle to transform the tilted projection back to its original shape (e.g., a rectangle). For each pixel in the projected image, the transformation matrix can be applied. By applying this transformation to all pixels in the projected image, the tilted projection can be corrected to approximate the original card shape, thus correcting the projection size to obtain the correct projection dimensions.

[0122] By implementing the embodiments of this disclosure, a first tilt angle of the reference object in the horizontal direction is determined based on the actual width and projected width of the reference object, and a second tilt angle of the reference object in the vertical direction is determined based on the actual length and projected length of the reference object. Based on the first and second tilt angles, the projection of the reference object in the facial image is corrected to update the projection size of the reference object in the facial image. In the case of a rectangular reference object being tilted, the two decomposed tilt angles can be conveniently calculated through the changes in the projection of the length and width, thereby correcting the projection and improving the accuracy of the reference object's projection and the accuracy of determining the relevant proportions based on the projection size, ultimately improving the accuracy of mask determination.

[0123] Optionally, the reference object is a rectangle; when there is a reference object in the facial image, the step of determining the target ratio between the projection size of the reference object in the facial image and the facial features includes:

[0124] In the case where there is a reference object in the facial image and the reference object is tilted, the facial image is identified to obtain the projected length and projected width of the reference object; the projected length and the projected width are both determined by the actual length, actual width and third tilt angle of the reference object;

[0125] The third tilt angle is calculated based on the actual length, actual width, projected length, and projected width of the reference object.

[0126] Based on the third tilt angle, the projection size of the reference object in the facial image is corrected, and a target ratio between the projection size and the facial features is determined.

[0127] In this embodiment of the disclosure, the tilt of the reference object does not need to be decomposed into tilt angles in two directions. When the user holds the reference object, the reference object can be regarded as rotating around a point in space by a specific angle, namely the third tilt angle.

[0128] Figure 7This is another schematic diagram of the tilt of the reference object provided in the embodiments of this disclosure; when the reference object and the face are on the same plane, the two opposite corner points are connected and extended into line 701; when the reference object and the face are on different planes, the two opposite corner points are connected and extended into line 702; the intersection of line 701 and line 702 can be considered as the point around which the reference object rotates; the included angle between line 701 and line 702 is the third tilt angle.

[0129] Similarly, after rotation, the length and width of the bank card will change. Assume the bank card is rotated around its normal direction (perpendicular to the plane) by a third tilt angle θ, and the projected length and width obtained from the acquired image are H' and W', respectively. Their ratio is... It will change. After rotation, the length and width dimensions of the bank card projected onto the plane can be calculated using the following geometric formula:

[0130] H ′ =H cosθ + Wsinθ

[0131] W ′ =Wcosθ + L sinθ

[0132] Calculate the aspect ratio after rotation, and the aspect ratio r′ of the projection:

[0133]

[0134] Substitute the rotated length and width into the formula:

[0135]

[0136] By calculating the aspect ratio r' of the rotated projection, we can inversely deduce the rotation angle θ. Substituting H = rW into the formula, we obtain:

[0137]

[0138] Simplified, we get:

[0139]

[0140] Solve this equation to obtain the third tilt angle θ. The solution can be obtained by numerical methods, such as using the bisection method or Newton's method, or by using image processing techniques to fit the equation and find the most suitable angle.

[0141] Similarly, based on the third tilt angle, the projection of the reference object in the facial image is corrected to update the projection size of the reference object in the facial image.

[0142] In implementing embodiments of this disclosure, a facial image is identified to obtain the projected length and width of a reference object. Given that the true length, true width, projected length, and projected width of the reference object are known, a third tilt angle is calculated. Based on this third tilt angle, the projection of the reference object in the facial image is corrected to update the projected size of the reference object in the facial image. By calculating the third tilt angle of rotation through changes in the projected length and width when the rectangular reference object is tilted, the projection can be corrected, thereby improving the accuracy of the reference object and ultimately improving the accuracy of mask determination.

[0143] Figure 8 This disclosure provides a mask determining device, device 80 including:

[0144] The image recognition module 801 is used to recognize the user's facial image to obtain facial features in the facial image; each facial feature represents the distance between two feature points in the facial image;

[0145] The reference ratio module 802 is used to determine, when there is a reference object in the facial image, a target ratio between the projection size of the reference object in the facial image and the facial features, and to determine that the target ratio is a first type; the first type represents the size relationship between the projection size and the facial features.

[0146] The facial proportion module 803 is used to determine a target proportion between two facial features when there is no reference object in the facial image, and to determine the target proportion as a second type; the second type represents the size relationship between the two facial features.

[0147] The mask determination module 804 is used to determine a mapping relationship corresponding to the proportion type of the target proportion, and to determine the target mask specification corresponding to the target proportion in the mapping relationship; different proportion types correspond to different mapping relationships; the mapping relationship is the correspondence between proportion and mask specification.

[0148] Optionally, the mask determination module includes:

[0149] A multi-ratio submodule is used to determine a set of mapping relationships corresponding to the ratio types of the target ratios when there are at least two target ratios; the ratio types of the at least two target ratios are either both of the first type or both of the second type; the set of mapping relationships includes at least two mapping relationships;

[0150] Each submodule is defined to determine the target mask specifications in different mapping relationships for different target proportions, so as to obtain at least two target mask specifications.

[0151] Optionally, there are two target proportions, both of which are of the first type; different mapping relationships in the mapping relationship set correspond to different facial features;

[0152] The sub-modules are defined separately, including:

[0153] The first specification unit is used to determine the first facial feature corresponding to the first target ratio and the first mapping relationship corresponding to the first facial feature, and in the first mapping relationship, to determine the first target mask specification corresponding to the first target ratio.

[0154] The second specification unit is used to determine the second facial feature corresponding to the second target ratio and the second mapping relationship corresponding to the second facial feature, and in the second mapping relationship, to determine the second target mask specification corresponding to the second target ratio.

[0155] The feature directions of the first facial feature and the second facial feature are perpendicular to each other; the feature direction is determined by the line connecting the two feature points.

[0156] Optionally, there are two target proportions, both of which are of the second type; different mapping relationships in the mapping relationship set correspond to different facial features;

[0157] The sub-modules are defined separately, including:

[0158] The third specification unit is used to determine the first facial feature and the second facial feature corresponding to the first target ratio, as well as the first mapping relationship corresponding to the first facial feature and the second facial feature, and to determine the first target mask specification corresponding to the first target ratio in the first mapping relationship.

[0159] The fourth specification unit is used to determine the third and fourth facial features corresponding to the second target ratio, as well as the second mapping relationship corresponding to the third and fourth facial features, and to determine the second target mask specification corresponding to the second target ratio in the second mapping relationship.

[0160] Wherein, the feature directions of the first facial feature and the second facial feature are perpendicular to each other; the feature directions of the third facial feature and the fourth facial feature are perpendicular to each other.

[0161] Optionally, the distance represented by each facial feature is either a horizontal or vertical distance; the facial proportion module includes:

[0162] The feature selection submodule is used to determine two target facial features with different feature directions from multiple facial features in the facial image; the two feature directions include a horizontal direction corresponding to the horizontal distance and a vertical direction corresponding to the vertical distance;

[0163] The feature region module is used to determine a reference facial feature among two target facial features based on a preset stability parameter for each facial feature; the features other than the reference facial feature are the contrast facial features.

[0164] The distance ratio submodule is used to determine the target ratio between two target facial features based on the ratio between the first distance corresponding to the reference facial feature and the second distance corresponding to the comparison facial feature.

[0165] Optionally, the reference object is a rectangle; the reference scale module includes:

[0166] The first tilting submodule is used to determine a first tilt angle of the reference object in the horizontal direction based on the actual width of the reference object and its projected width in the facial image when there is a reference object in the facial image and the reference object is tilted.

[0167] The second tilting submodule is used to determine the second tilt angle of the reference object in the vertical direction based on the actual length of the reference object and the projected length in the facial image;

[0168] The first correction submodule is used to correct the projection size of the reference object in the facial image based on the first tilt angle and the second tilt angle, and to determine the target ratio between the projection size and the facial features.

[0169] Optionally, the reference object is a rectangle; the reference scale module includes:

[0170] The parameter relationship submodule is used to identify the facial image to obtain the projection length and projection width of the reference object when the reference object is tilted; the projection length and projection width are both determined by the actual length, actual width and third tilt angle of the reference object;

[0171] The third tilting submodule is used to calculate the third tilting angle based on the actual length, actual width, projected length, and projected width of the reference object.

[0172] The second correction submodule is used to correct the projection size of the reference object in the facial image based on the third tilt angle, and to determine the target ratio between the projection size and the facial features.

[0173] In summary, in this embodiment of the present disclosure, facial features are obtained by recognizing a user's facial image. Each facial feature represents the distance between two feature points in the facial image. When there is a reference object in the facial image, the target ratio between the projection size of the reference object in the facial image and the facial features is determined, and the target ratio is determined to be of the first type. When there is no reference object in the facial image, the target ratio between two facial features is determined, and the target ratio is determined to be of the second type. The corresponding target ratio can be obtained under different conditions with or without a reference object, and the target mask specification indicated by the target ratio is determined in the mapping relationship corresponding to the ratio type. The corresponding target mask specification can be determined directly based on the corresponding target ratio, avoiding the tedious calculation process involving the measurement values, sizes, image scaling and other contents of different facial features, thus improving the efficiency of mask determination. At the same time, since the mask is determined in real time based on the actual ratio between features in the facial image, rather than using the fixed preset size of a specific facial feature as a reference, the mask determination process also has high flexibility, thereby improving the accuracy of mask determination.

[0174] This application also provides an electronic device, such as... Figure 9 As shown, it includes a processor 1001, a communication interface 1002, a memory 1003, and a communication bus 1004, wherein the processor 1001, the communication interface 1002, and the memory 1003 communicate with each other through the communication bus 1004.

[0175] Memory 1003 is used to store computer programs.

[0176] When the processor 1001 executes the program stored in the memory 1003, it implements the steps in the mask determination method described above, which will not be repeated here.

[0177] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not indicate that there is only one bus or one type of bus.

[0178] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0179] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0180] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be 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, or discrete hardware components.

[0181] In another embodiment provided in this application, a computer-readable storage medium is also provided, which stores instructions that, when executed on a computer, cause the computer to perform the mask determination method described in the above embodiments.

[0182] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to perform the mask determination method described in the above embodiments.

[0183] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0184] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0185] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. For embodiments of devices, electronic devices, computer-readable storage media, and computer program products containing instructions, the descriptions are relatively simple because they are basically similar to the method embodiments; relevant parts can be referred to the descriptions of the method embodiments.

[0186] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application are included within the scope of protection of this application.

Claims

1. A method for determining a face mask, characterized in that, The method includes: Recognize the user's facial image to obtain facial features in the facial image; each facial feature represents the distance between two feature points in the facial image; In the case of a reference object in the facial image, a target ratio between the projection size of the reference object in the facial image and the facial features is determined, and the target ratio is determined to be of a first type; the first type represents the size relationship between the projection size and the facial features. In the absence of a reference object in the facial image, a target ratio between two facial features is determined, and the target ratio is identified as a second type; the second type represents the size relationship between the two facial features. Determine the mapping relationship corresponding to the proportion type of the target proportion, and determine the target mask size corresponding to the target proportion in the mapping relationship; different proportion types correspond to different mapping relationships; the mapping relationship is the correspondence between proportion and mask size.

2. The method according to claim 1, characterized in that, The step of determining the mapping relationship corresponding to the proportion type of the target proportion, and determining the target mask size corresponding to the target proportion in the mapping relationship, includes: When there are at least two target proportions, a set of mapping relationships corresponding to the proportion types of the target proportions is determined; the proportion types of at least two target proportions are either both of the first type or both of the second type; the set of mapping relationships includes at least two mapping relationships. For different target proportions, the target mask specifications are determined in different mapping relationships to obtain at least two target mask specifications.

3. The method according to claim 2, characterized in that, There are two target proportions, both of which are of the first type; different mapping relationships in the mapping relationship set correspond to different facial features; The step of determining the target mask specifications in different mapping relationships for different target proportions to obtain two target mask specifications includes: Determine the first facial feature corresponding to the first target ratio and the first mapping relationship corresponding to the first facial feature, and determine the first target mask specification corresponding to the first target ratio in the first mapping relationship; Determine the second facial feature corresponding to the second target ratio, and the second mapping relationship corresponding to the second facial feature, and in the second mapping relationship, determine the second target mask specification corresponding to the second target ratio; The feature directions of the first facial feature and the second facial feature are perpendicular to each other; the feature direction is determined by the line connecting the two feature points.

4. The method according to claim 2, characterized in that, There are two target proportions, both of which are of the second type; different mapping relationships in the mapping relationship set correspond to different facial features; The step of determining the target mask specifications in different mapping relationships for different target proportions to obtain two target mask specifications includes: Determine the first facial feature and the second facial feature corresponding to the first target ratio, as well as the first mapping relationship corresponding to the first facial feature and the second facial feature, and determine the first target mask specification corresponding to the first target ratio in the first mapping relationship; Determine the third and fourth facial features corresponding to the second target ratio, and the second mapping relationship corresponding to the third and fourth facial features, and in the second mapping relationship, determine the second target mask specification corresponding to the second target ratio; Wherein, the feature directions of the first facial feature and the second facial feature are perpendicular to each other; the feature directions of the third facial feature and the fourth facial feature are perpendicular to each other.

5. The method according to claim 1, characterized in that, Each facial feature represents a horizontal or vertical distance; the step of determining the target ratio between two facial features in the absence of a reference object in the facial image includes: Among multiple facial features in the facial image, two target facial features with different feature directions are identified; the two feature directions include a horizontal direction corresponding to the horizontal distance and a vertical direction corresponding to the vertical distance. In the two target facial features, a reference facial feature is determined based on a preset stability parameter for each facial feature; features other than the reference facial feature are contrast facial features. The target ratio between the two target facial features is determined based on the ratio between the first distance corresponding to the reference facial feature and the second distance corresponding to the comparison facial feature.

6. The method according to claim 1, characterized in that, The reference object is rectangular; when there is a reference object in the facial image, the step of determining the target ratio between the projection size of the reference object in the facial image and the facial features includes: In the case where there is a reference object in the facial image and the reference object is tilted, the first tilt angle of the reference object in the horizontal direction is determined based on the actual width of the reference object and the projected width in the facial image. The second tilt angle of the reference object in the vertical direction is determined based on the actual length of the reference object and its projected length in the facial image; Based on the first tilt angle and the second tilt angle, the projection size of the reference object in the facial image is corrected, and a target ratio between the projection size and the facial features is determined.

7. The method according to claim 1, characterized in that, The reference object is rectangular; when there is a reference object in the facial image, the step of determining the target ratio between the projection size of the reference object in the facial image and the facial features includes: In the case where there is a reference object in the facial image and the reference object is tilted, the facial image is identified to obtain the projected length and projected width of the reference object; the projected length and the projected width are both determined by the actual length, actual width and third tilt angle of the reference object; The third tilt angle is calculated based on the actual length, actual width, projected length, and projected width of the reference object. Based on the third tilt angle, the projection size of the reference object in the facial image is corrected, and a target ratio between the projection size and the facial features is determined.

8. A mask determining device, characterized in that, The device includes: An image recognition module is used to recognize a user's facial image to obtain facial features in the facial image; each facial feature represents the distance between two feature points in the facial image; The reference ratio module is used to determine, when there is a reference object in the facial image, the target ratio between the projection size of the reference object in the facial image and the facial features, and to determine that the target ratio is a first type; the first type represents the size relationship between the projection size and the facial features. A facial proportion module is used to determine a target proportion between two facial features when there is no reference object in the facial image, and to determine the target proportion as a second type; the second type represents the size relationship between the two facial features. The mask determination module is used to determine a mapping relationship corresponding to the proportion type of the target proportion, and to determine the target mask specification corresponding to the target proportion in the mapping relationship; different proportion types correspond to different mapping relationships; the mapping relationship is the correspondence between proportion and mask specification.

9. An electronic device, characterized in that, include: The system includes a processor, a communication interface, a memory, and a communication bus; the processor, communication interface, and memory communicate with each other via the communication bus. Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the steps of the mask determination method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the mask determination method as described in any one of claims 1 to 7.

Images