An exposure control method and electronic device

By determining the brightness data of the target reference area and the facial target area, and using their relative relationship to dynamically adjust the exposure parameters, the problem of the inability to accurately perceive facial reflection characteristics in traditional exposure control methods is solved, thus realizing personalized exposure control and improving image quality.

CN123123728APending Publication Date: 2026-11-17LENOVO (BEIJING) LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611132685.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-28
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Traditional exposure control methods cannot accurately perceive the facial reflection characteristics of a subject's face, leading to exposure deviations and affecting image quality.

Method used

By determining the brightness data of the target reference area and the facial target area, the exposure parameters are dynamically adjusted using their relative relationship. In particular, the visible area of ​​the sclera or the accessory area is used as a reference. Combined with a preset mapping relationship table, the target exposure brightness is calculated and the exposure parameters of the image acquisition device are controlled.

Benefits of technology

It enables personalized exposure control based on individual facial reflection characteristics, improving image quality, avoiding overexposure or underexposure, and enhancing the overall image effect.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123123728A_ABST
    Figure CN123123728A_ABST
Patent Text Reader

Abstract

This disclosure provides an exposure control method and an electronic device, relating to the field of image acquisition technology. The exposure control method includes: acquiring an image containing a target face; determining brightness data of a target reference region and a facial target region in the image; determining a target exposure brightness corresponding to the target face based on the relative relationship between the brightness data of the target reference region and the facial target region; and controlling the exposure parameters of the image acquisition device using the target exposure brightness.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This disclosure relates to the field of image acquisition technology, and in particular to an exposure control method and electronic device. Background Technology

[0002] In scenarios involving automatic face exposure, traditional solutions struggle to adjust the exposure value to a suitable level that corresponds to the current user's facial features. The algorithms fail to accurately perceive the facial reflection characteristics of the subject, leading to exposure deviations and thus compromising the overall image quality. Summary of the Invention

[0003] This disclosure provides an exposure control method and an electronic device.

[0004] According to a first aspect of this disclosure, an exposure control method is provided, comprising: Obtain an image containing the target face; Determine the brightness data of the target reference region and the facial target region in the image; Based on the relative relationship between the brightness data of the target reference area and the facial target area, the target exposure brightness corresponding to the target face is determined. The exposure parameters of the image acquisition device are controlled using the target exposure brightness.

[0005] According to a second aspect of this disclosure, an electronic device is provided, including a device body and an image acquisition device disposed on the device body, wherein the image acquisition device includes an image sensor, and the device body further includes a signal processor; The image sensor is used to: acquire images containing the target human face; The signal processor is configured to: acquire the image containing the target face from the image sensor; determine the brightness data of the target reference region and the facial target region in the image, and determine the relative relationship between the brightness data of the target reference region and the facial target region; based on the relative relationship, determine the target exposure brightness corresponding to the target face, and use the target exposure brightness to control the exposure parameters of the image acquisition device.

[0006] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0007] The above and other objects, features, and advantages of this disclosure will become readily apparent from the following detailed description of exemplary embodiments, taken in conjunction with the accompanying drawings. Several embodiments of this disclosure are illustrated in the drawings by way of example and not limitation, in which: In the accompanying drawings, the same or corresponding reference numerals indicate the same or corresponding parts.

[0008] Figure 1 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 1 ; Figure 2 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 2 ; Figure 3 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 3 ; Figure 4 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 4 ; Figure 5 A schematic diagram of a scenario illustrating an exposure control method according to an embodiment of this disclosure is shown; Figure 6 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 5 ; Figure 7 A schematic diagram of the composition structure of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0009] To make the objectives, features, and advantages of this disclosure more apparent and understandable, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort are within the scope of protection of this disclosure.

[0010] In modern camera systems (especially smartphones, video conferencing equipment, and security monitoring), there is a fundamental flaw in face-priority automatic exposure algorithms that has long remained unresolved: the algorithm cannot accurately perceive the facial reflection characteristics of the subject.

[0011] Traditional exposure control methods work by first detecting the face area in the image, and then adjusting the brightness of that face area to a preset ideal intermediate value (typically corresponding to medium facial reflectivity). This leads to the following problems: When the surface of a face has low reflectivity: in order to increase the exposure brightness to the middle value, the overall exposure will be excessively increased, resulting in severe overexposure of the background and other objects, loss of detail, and image distortion.

[0012] When the face has high reflectivity: in order to reduce the exposure brightness to the middle value, the overall exposure will be reduced, making the picture dark, the background underexposed, and the person's complexion poor.

[0013] The fundamental reason is that image sensors and processors can only measure the light intensity (brightness) reaching the sensor, but cannot distinguish between "the object itself being dark" and "insufficient lighting." Traditional algorithms push all faces "equally" towards the same brightness target, ignoring the huge natural differences in the facial reflective characteristics of different faces, thus damaging the overall image quality.

[0014] Some existing improvement schemes attempt to indirectly infer facial reflex characteristics by recognizing facial biometric features (such as nose bridge height and lip shape), but such methods have serious drawbacks such as low accuracy and high ethical risks, making them difficult to apply on a large scale.

[0015] Figure 1 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 1 ,like Figure 1 As shown, an exposure control method includes: Step S101: Obtain an image containing the target face.

[0016] In this embodiment, an image of the target face can be acquired using an image acquisition device. This image can be a single static image or a frame from a continuous video stream. The image contains at least one target face. If the image contains multiple faces, some or all of the faces can be identified as the target face based on the proportion of each face in the image and facial occlusions. For example, if the proportion of a face in the image is less than a certain threshold, the face can be considered a background person and not identified as a target face; if a face is largely covered by an occlusion such as a mask or hair, resulting in a very small effective facial area, it can also be excluded from being identified as a target face.

[0017] Step S102: Determine the brightness data of the target reference area and the facial target area in the image.

[0018] In this embodiment, after acquiring an image containing the target face, it is necessary to determine the target reference region and the facial target region from the image. The target reference region refers to a physiological region on the target face with stable optical properties. The brightness difference between individuals in this region is less than a very small threshold, meaning the brightness is almost the same across individuals. Furthermore, it maintains relatively stable reflective properties under different external lighting conditions, thus serving as a brightness reference throughout the exposure control process. The facial target region refers to the facial skin region on the target face excluding the target reference region. The brightness of this region varies significantly between individuals. In one example, the target reference region could be the sclera (white of the eye), teeth, or accessories such as sunglasses. Since accessories like sunglasses have higher variability than the sclera and teeth, if the target reference region is an accessory, the relative relationship needs to be corrected after determining it to mitigate the error interference from accessories. The facial target region refers to the usable facial skin region on the target face excluding the target reference region, and after removing abnormal areas such as the eyes, eyebrows, lips, and shadows / highlights.

[0019] In one implementation, the target face is first located in the image using a face detection algorithm to obtain the face bounding box. Then, a target reference region and a facial target region are located within the target face region. After determining the target reference region and the facial target region, luminance data extraction is performed on the pixels in these two regions. For example, the image is converted from the RGB color space to a color space that includes a luminance channel, such as the Y channel in the YCbCr color space or the L channel in the Lab color space. Then, the pixel values ​​of all pixels in the target reference region and the facial target region in the luminance channel are read respectively. Based on the read pixel values, the luminance data of the target reference region and the facial target region in the image are determined.

[0020] Step S103: Based on the relative relationship between the brightness data of the target reference area and the facial target area, determine the target exposure brightness corresponding to the target face.

[0021] In this embodiment, after obtaining the brightness data of the target reference area and the brightness data of the facial target area, it is also necessary to determine the relative relationship between the two. This relative relationship is used to characterize the brightness of the facial target area relative to the target reference area. Since the target reference area has highly consistent brightness characteristics across different individuals, the magnitude of this relative relationship actually reflects the reflectivity of the facial target area to light, i.e., the actual brightness characteristics of the facial target area. This relative relationship can be the difference or ratio between the brightness data of the target reference area and the facial target area, etc. The magnitude of this relative relationship is less affected by ambient light, because when ambient light changes, the brightness of the target reference area and the facial target area will change in the same direction with the ambient light.

[0022] In this embodiment, the target exposure brightness corresponding to the target face can be determined based on a relative relationship. The target exposure brightness can be dynamically determined according to the actual reflective characteristics of the current target face. In one example, the calculated relative relationship can be matched with a preset numerical range, with different numerical ranges corresponding to different target exposure brightness; or, through a pre-defined functional relationship, the relative relationship value can be used as the input variable, and the corresponding target exposure brightness can be obtained through function calculation.

[0023] Step S104: Use the target exposure brightness to control the exposure parameters of the image acquisition device.

[0024] In this embodiment, after determining the target exposure brightness corresponding to the target face, the exposure parameters of the image acquisition device are adjusted and controlled using this target exposure brightness. In one example, a corresponding exposure parameter adjustment signal can be generated based on the target exposure brightness, and this signal is sent to the drive control circuit of the image acquisition device, which then changes the exposure parameters of the device. The exposure parameters of the image acquisition device include at least one of the following: exposure time of the image sensor, gain (including analog and digital gain), lens aperture parameters, fill light parameters, multi-frame exposure strategy, and HDR (High Dynamic Range) exposure combination parameters.

[0025] In this disclosure, an image containing a target face is acquired, brightness data of a target reference region and a facial target region are determined, a target exposure brightness is determined based on the relative relationship between the two, and the exposure parameters of the image acquisition device are controlled using this target exposure brightness. The target reference region exhibits highly consistent brightness characteristics across different individuals. The relative relationship determined based on the brightness data of the target reference region and the facial target region objectively reflects the brightness differences in the facial regions of different individuals, thereby setting a precisely corresponding target exposure brightness for each individual and avoiding the problem of unsatisfactory exposure results in traditional algorithms.

[0026] Figure 2 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 2 ,like Figure 2 As shown, the target reference area can be the visible area of ​​the sclera of the target face. One exposure control method includes: Step S201: Obtain an image containing the target face.

[0027] The specific implementation details of step S201 are similar to those of step S101, and will not be repeated here.

[0028] Step S202: Determine the brightness data of the visible sclera area and the target facial area of ​​the target face.

[0029] In this embodiment, the target reference area can be the visible sclera region of the target face. The sclera is the outermost layer of the eyeball wall, located around the iris and pupil, also known as the "white of the eye." The color and brightness of the sclera vary very little between individuals, and under normal healthy conditions, it presents a stable white or pale white appearance. In one example, in an image containing the target face, the eye region and facial target region of the target face can be determined first based on a facial landmark detection algorithm. Within the eye region, an image segmentation algorithm can be used to distinguish the sclera region from the iris region and pupil region, thereby obtaining the visible sclera region of the target face. After determining the visible sclera region and facial target region of the target face, it is also necessary to extract the brightness data of these two regions, for example, by reading the pixel values ​​of each pixel in the brightness channel within these two regions, thereby obtaining the brightness data of the visible sclera region and the facial target region.

[0030] Step S203: Based on the relative relationship between the brightness data of the visible area of ​​the sclera and the target area of ​​the face, determine the target exposure brightness corresponding to the target face.

[0031] In this embodiment, the target exposure brightness can also be determined based on the relative relationship between the brightness data of the visible scleral area and the brightness data of the facial target area. That is, the brightness data of the visible scleral area is used as a reference benchmark, and the brightness data of the facial target area is used as the comparison object to determine the relative relationship between the two. Then, the target exposure brightness corresponding to the target face is determined based on this relative relationship. For example, the relative relationship is matched with a preset numerical range, with different numerical ranges corresponding to different target exposure brightness; or the relative relationship value is used as an independent variable input through a preset functional relationship, and the corresponding target exposure brightness is obtained by function calculation.

[0032] Step S204: Use the target exposure brightness to control the exposure parameters of the image acquisition device.

[0033] The specific implementation details of step S204 are similar to those of step S104, and will not be repeated here.

[0034] In this disclosure, the physiological characteristic of minimal brightness difference in the visible sclera region of the target face among different individuals is used as a reference benchmark. This effectively eliminates the interference of ambient light variations on exposure determination, making the calculation of relative relationships more accurate and reliable. Determining the target exposure brightness by the relative brightness relationship between the visible sclera region and the target facial region provides differentiated exposure schemes for individuals with different reflectivity characteristics, effectively improving image quality.

[0035] Figure 3 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 3 ,like Figure 3 As shown, an exposure control method includes: Step S301: Obtain an image containing the target face.

[0036] Step S302: Determine the brightness data of the visible sclera area and the target facial area of ​​the target face.

[0037] The specific implementation details of steps S301-S302 are similar to those of steps S201-S202, and will not be repeated here.

[0038] Step S303: Determine the first average brightness of the visible area of ​​the sclera and the second average brightness of the target facial area.

[0039] In this embodiment, after determining the brightness data of the visible scleral region and the facial target region of the target face, it is also necessary to calculate the average brightness of the visible scleral region and the facial target region respectively. In one example, after extracting the brightness channel values ​​of each pixel in the visible scleral region, the average brightness value of these pixels is calculated, and this average value is recorded as the first average brightness. Similarly, after extracting the brightness channel values ​​of each pixel in the facial target region, the average brightness value of these pixels is calculated, and this average value is recorded as the second average brightness.

[0040] In one example, before calculating the average, the pixel brightness values ​​in the visible area of ​​the sclera and the target area of ​​the face can be preprocessed. For example, median filtering can be used to remove extreme pixel values ​​caused by local reflections or shadows in the area, or abnormal pixel values ​​located at the top and bottom of the brightness value distribution can be removed. Then, the average value of the remaining pixel values ​​can be calculated, thereby further improving the stability and representativeness of the average brightness value.

[0041] In this embodiment, to improve the accuracy of the brightness of the visible scleral region, the brightness information of the visible scleral regions of the left and right eyes can be used to determine the final first average brightness. For example, when the target face includes both left and right eyes, the brightness of the visible scleral regions of the left and right eyes is extracted separately, and the average brightness of the visible scleral regions of the left and right eyes is calculated. In one example, the average brightness of the left and right eyes can be arithmetically averaged, and the resulting average value can be used as the first average brightness of the visible scleral region. In another example, the quality of the visible scleral regions of the left and right eyes can be evaluated separately. For example, the proportion of effective pixels in each region can be calculated. Effective pixels refer to pixels whose brightness values ​​are within the normal range and conform to the scleral color characteristics. Different weights are assigned to the average brightness of the left and right eyes based on the proportion of effective pixels, with a higher weight assigned to the side with a higher proportion of effective pixels and a lower weight assigned to the side with a lower proportion of effective pixels. Then, a weighted average is performed to obtain the first average brightness.

[0042] In another example, the average brightness of the left and right sclera can be compared. If the difference is less than a preset threshold, it indicates good brightness consistency between the two sclera, and the average of the two values ​​is used as the reference brightness. If the difference is greater than or equal to the preset threshold, it indicates abnormalities such as uneven illumination on one side (e.g., one side is illuminated while the other is backlit), one eye is obstructed (e.g., by hair, eye mask, or sunglasses), or glare from one eyeglass lens. In this case, the brightness value with the higher value or the higher percentage of effective pixels is selected as the reference brightness to avoid contamination of the brightness value by low-quality brightness data from the obstructed or glaring side. This ensures the accuracy and stability of the first average brightness, thereby ensuring the reliability of the facial reflection characteristics calculated based on this first average brightness and avoiding exposure control deviations caused by unilateral abnormalities.

[0043] Step S304: Determine the facial reflection characteristics corresponding to the target face based on the ratio between the second average brightness and the first average brightness.

[0044] In this embodiment, it is also necessary to determine the ratio of the second average brightness to the first average brightness. This ratio characterizes the reflectivity of the target face surface relative to the scleral surface. If the ratio is small, it indicates that the brightness of the target facial area is much lower than the brightness of the visible scleral area; if the ratio is large, it indicates that the brightness of the target facial area is much higher than the brightness of the visible scleral area; if the ratio is close to 1, it indicates that the brightness of the target facial area is close to the brightness of the visible scleral area. Since the reflectivity of the scleral surface in the human body is highly consistent among different individuals, this ratio directly corresponds to the reflectivity of the target facial area. Therefore, this ratio can be directly determined as the facial reflectivity of the target face.

[0045] In one example, the ratio can be corrected to obtain the facial reflection characteristics corresponding to the target face. The sclera region may be affected by factors such as partial occlusion, eyeglass reflection, eye fatigue and congestion, or conjunctivitis, causing the first average brightness to deviate from the normal value; or the facial target area may be affected by factors such as local shadows, highlights, and cosmetic reflection, causing the second average brightness to deviate from the normal value. In these cases, the ratio of the second average brightness to the first average brightness can be corrected. For example, it can be corrected based on the size comparison of the facial target area. When the target face occupies too small a proportion in the image and the number of pixels in the facial target area is small, the reliability of the statistical second average brightness is reduced. In this case, a lower confidence weight can be assigned to the calculated ratio, so that the final facial reflection characteristics tend to be at a moderate level, avoiding extreme bias caused by small sample statistics.

[0046] Step S305: Determine the target exposure brightness for the target face from the target mapping table based on the facial reflection characteristics.

[0047] In this embodiment, after determining the facial reflection characteristics of the target face, the target exposure brightness for the target face is determined by searching a pre-established target mapping table based on these characteristics. The target mapping table records the correspondence between different facial reflection characteristics and their corresponding optimal exposure brightness. This mapping table can be pre-established through large-scale data calibration. For example, a large amount of sample data with different facial reflection characteristics can be collected under standard lighting conditions. The ratio is calculated based on the first and second average brightness of each sample, and an optimal exposure brightness is determined for each sample to achieve the best facial imaging effect. The correspondence between the ratio range and the optimal exposure brightness value is stored in tabular form. In the target mapping table, different facial reflection characteristics (usually represented by ratio ranges) correspond to different target exposure brightness. For example, when the ratio corresponding to the facial reflection characteristic is small (low facial reflectivity), the target exposure brightness in the target mapping table is lower than the standard intermediate gray value to avoid over-increasing brightness; when the ratio is medium (medium facial reflectivity), the corresponding target exposure brightness is close to the standard intermediate gray value; when the ratio is large (high facial reflectivity), the corresponding target exposure brightness is higher than the standard intermediate gray value to avoid insufficient brightness.

[0048] Step S306: Control the exposure parameters of the image acquisition device using the target exposure brightness.

[0049] The specific implementation details of step S306 are similar to those of step S204, and will not be repeated here.

[0050] In this disclosure, a first average brightness of the visible scleral region and a second average brightness of the facial target region are determined, and the ratio between the two is calculated to determine the facial reflectivity. Then, the target exposure brightness is determined using a target mapping table. Therefore, using the average value as the statistical characteristic value of the brightness data is simple and efficient. The facial reflectivity obtained through ratio calculation is insensitive to changes in external light intensity, thus eliminating the need for parameter adjustments for different lighting environments. Furthermore, the target exposure brightness is quickly obtained by looking up a pre-established target mapping table, resulting in low computational overhead and low processing latency, meeting the needs of real-time applications such as video streaming. Simultaneously, different facial reflectivity characteristics in the mapping table correspond to different target exposure brightnesses, allowing the exposure target to adaptively match the actual reflectivity characteristics of different individuals, thereby achieving personalized exposure control.

[0051] In another embodiment, step S202, "determining the brightness data of the visible scleral area and the target facial area of ​​the target face," includes: Keypoint detection is performed on the image to determine the region of interest (ROI) of the eye based on the detected eye keypoints, and the visible area of ​​the sclera is determined within the ROI. The facial target region is obtained by excluding at least one of the eye region of interest, mouth region, hair region, and accessory region from the facial region of the target face. The pixel brightness values ​​in the visible area of ​​the sclera and the target area of ​​the face are filtered and / or outlier removed to obtain the corresponding brightness data.

[0052] In this embodiment, the brightness data of the visible scleral region and the target facial region of the target face are determined. Before this, the specific locations of the visible scleral region and the target facial region need to be determined, which requires performing facial landmark detection on the image containing the target face. Facial landmark detection is a computer vision technique that can locate the position coordinates of semantically labeled facial feature points in a face image. These feature points include, but are not limited to, eye contour points, eyebrow contour points, nose tip, lip contour points, and facial contour points. In this embodiment, the facial landmark detection algorithm is used to locate the contour key points of the left and right eyes within the facial area of ​​the target face. These key points are distributed at the edges of the upper and lower eyelids and the inner and outer corners of the eyes, accurately outlining the shape boundaries of the eyes. After detecting the eye key points, the region of interest (ROI) is determined in the image based on the position coordinates of these key points. The ROI refers to a local image region containing the complete eye structure (including the sclera, iris, pupil, and upper and lower eyelids). After determining the ROI, the visible scleral region is further determined within this region. The sclera is the outermost layer of the eyeball wall, located around the iris and pupil, and appears white or off-white. First, based on a geometric model, the region of interest (ROI) of the eye can be divided into estimated regions for the pupil, iris, and sclera (white of the eye). Since the sclera's color is clearly distinguishable from the iris, pupil, and eyelids in color space, a color thresholding method can be used to extract the corresponding pixels from the estimated sclera region. For example, the pixels in the estimated sclera region can be converted from the RGB color space to the YCbCr or Lab color space. Based on the typical characteristics of the sclera in the color space (e.g., in the YCbCr space, the sclera has a high Y value, a low Cb value, and a low Cr value), a color threshold range is set, and pixels that meet this threshold range are marked as candidate sclera pixels. Then, connected component analysis is performed on these candidate pixels to remove noise regions with excessively small areas; the remaining connected components represent the visible sclera region.

[0053] In this embodiment, an exclusion operation is performed within the entire facial region of the target face to obtain a clean facial target region. The facial region of the target face refers to the area enclosed by the outer contour boundary of the face determined by the face detection algorithm or the facial keypoint detection algorithm, which covers the entire range of the face. However, this facial region includes various non-facial surface areas, which may interfere with the accuracy of the brightness data and therefore need to be excluded. For example, the region of interest for the eyes needs to be excluded from the facial region because the eye region includes the sclera, iris, and pupil. The sclera region has already been extracted separately as a target reference region, and the brightness of the iris and pupil is not comparable to the brightness of the skin. Therefore, the entire region of interest for the eyes should be excluded from the facial target region. The mouth region also needs to be excluded because the color and reflective properties of the lips also differ significantly from the facial surface. Hair regions also need to be excluded, including eyebrows, hair, sideburns, and beards, as the brightness characteristics of these hair regions differ from the facial surface. Accessory regions also need to be excluded, such as areas covered by eyeglass frames, masks, lip studs, forehead ornaments, etc. After the above exclusion process, the remaining part of the facial area is the pure facial target area, which represents the effective facial surface area of ​​the target face that can be used for brightness analysis.

[0054] In this embodiment, after determining the visible scleral area and the facial target area, preprocessing is required on the pixel brightness values ​​within these two areas to obtain accurate and reliable brightness data. Preprocessing may include filtering and / or outlier removal. Filtering smooths the distribution of brightness values ​​within the area, reducing brightness fluctuations caused by image noise or local fine textures. In one implementation, median filtering can be used to process the pixel brightness values ​​within the area. Median filtering effectively removes impulse noise and salt-and-pepper noise while preserving edge information of brightness changes. In another implementation, mean filtering can be used to smooth the pixel brightness values ​​within the area. Mean filtering calculates the average brightness of all pixels within the filtering window to replace the brightness value of the center pixel, effectively suppressing Gaussian noise. Outlier removal removes extreme brightness values ​​caused by local highlights, shadows, or occlusion edges within the area. In one implementation, the brightness values ​​of all pixels within the area are statistically analyzed, and the mean and standard deviation of the brightness values ​​are calculated. Pixels with brightness values ​​exceeding three times the standard deviation of the mean are marked as outliers and removed. In another implementation, the pixel brightness values ​​within the region can be sorted, and pixels with a certain percentage of brightness values ​​at the top and bottom of the distribution can be removed. For example, the top 5% and bottom 5% of pixels can be removed. The remaining pixel brightness values ​​are then further processed. In yet another implementation, filtering and outlier removal can be combined. Outlier removal is performed first to remove pixels with abnormally high or low brightness caused by lens reflections or shadows. Then, median filtering is performed to further smooth the brightness distribution and eliminate the influence of image noise. After these preprocessing steps, the brightness data corresponding to the visible scleral area and the facial target area are obtained.

[0055] In this embodiment, if the visible area of ​​the sclera is very small, such as when the target person is far from the camera and the brightness data of the visible area of ​​the sclera cannot be determined, a reference eye image (or a reference image with a large proportion of the face) corresponding to the current target face can be found in a preset sclera database based on the current target face. These reference images can be template images with scleral brightness values ​​pre-acquired and calibrated under standard lighting conditions. The brightness level of the reference image is adjusted to match the overall brightness level of the current image, even if the brightness statistics of the reference image (e.g., the average brightness of the entire image or the average brightness of the background area) are basically consistent with the corresponding brightness statistics of the current image. After the brightness adjustment is completed, the brightness of the sclera region in the adjusted reference image is used as the substitute brightness data of the visible area of ​​the sclera of the target face in the current image. Thus, even when the scleral brightness data cannot be extracted from the current image, a reliable scleral reference brightness can still be obtained, ensuring the availability and continuity of the exposure control method in various complex scenarios.

[0056] In this disclosure, keypoint detection is performed on the image to determine the region of interest (ROI) for the eyes and the visible region of the sclera. The facial target region is obtained by excluding the areas around the eyes, mouth, hair, and accessories. Furthermore, the pixel brightness values ​​of both regions are filtered and outlier removed. Thus, keypoint detection accurately extracts the visible region of the sclera and eliminates various interfering regions to obtain a clean facial target region. This avoids interference from brightness data from non-facial surface areas, effectively improving the accuracy of the comparison between the brightness data of the two regions. Simultaneously, filtering and outlier removal of the pixel brightness values ​​in both regions eliminates the influence of abnormal pixels such as image noise, lens reflections, and shadow edges, further improving the accuracy and stability of the brightness data. This makes the subsequent calculated relative relationships more reliable, thereby ensuring improved exposure control precision.

[0057] In another embodiment, the target reference region is the accessory region formed by the accessories worn by the target face. Based on the relative relationship between the target reference region and the brightness data of the facial target data, the target exposure brightness corresponding to the target face is determined, including: The brightness data of the accessory area is determined based on the accessory information obtained from the identified accessory area; Based on the relative relationship between the brightness data of the accessory area and the facial target area, the target exposure brightness corresponding to the target face is determined.

[0058] In this embodiment, the target reference region can be the accessory region formed by accessories worn on the target face. Accessories can be any identifiable object worn on the human face, such as earrings, necklaces, hair clips, glasses (including sunglasses), or forehead ornaments. When the accessory occupies a certain area in the image and its surface reflective properties are known or can be obtained through recognition, the accessory region can be used as the target reference region. First, an accessory recognition operation can be performed on the image to identify the accessory region worn on the target face and obtain the accessory information. Accessory region recognition can employ deep learning-based object detection methods, such as training an accessory detection model. This model can locate the bounding box of the accessory region in the face image and identify the category of the accessory, such as earrings, necklaces, glasses, or hair clips. After identifying the accessory area, further accessory information is obtained. This information includes the accessory's category (e.g., earrings, necklaces, eyeglass frames), material (e.g., metal, plastic, resin, fabric), color (e.g., gold, silver, black, white), surface treatment (e.g., polished, matte, brushed), and model number (e.g., product code for a specific brand or series). After obtaining this information, the brightness data of the accessory area is determined. One method is to directly extract the brightness value of the accessory area from the image. This involves reading the brightness values ​​of all pixels within the accessory area's bounding box on the brightness channel and calculating their average or weighted average as the brightness data for the accessory area. Another method is to first obtain the known standard brightness value of the accessory under standard lighting conditions based on the accessory information. Then, this standard brightness value is adjusted according to the overall brightness level of the current image to adapt to the current lighting conditions, and the adjusted value is used as the brightness data for the accessory area.

[0059] In this embodiment, after determining the brightness data of the accessory area (i.e., the target reference area) and the brightness data of the facial target area, the relative relationship between the two is calculated. This relative relationship can be a ratio obtained by dividing the brightness data of the facial target area by the brightness data of the accessory area. Since the accessory area has known or identifiable reflective properties, its brightness level in the image is affected by both the inherent reflectivity of the accessory itself and the current ambient light. The brightness of the facial target area is simultaneously affected by the reflectivity of the face surface and the same ambient light. Therefore, by calculating the ratio between the two, the brightness difference between the facial target area and the accessory area under the same ambient light can objectively reflect the relative relationship between the facial reflective properties of the target face and the reflective properties of the accessory. After determining this relative relationship, the target exposure brightness can also be determined from the target mapping table based on this relationship. Here, the target mapping table needs to set different mapping relationships for different accessory types because different accessories have different inherent reflectivities, and the same ratio corresponds to different facial reflective properties under different accessory types. For example, the same ratio may represent completely different facial reflective properties when gold earrings are used as accessories and when white pearls are used as accessories. Therefore, it is necessary to pre-define the mapping relationship table corresponding to each type of accessory.

[0060] In this disclosure, when the scleral region cannot be stably extracted due to conditions such as closed eyes, eye obstruction, or wearing sunglasses, the accessory region provides an alternative reference scheme, ensuring the usability and continuity of the exposure control method in various complex scenarios. By identifying the accessory and obtaining information such as the category, material, and color of the accessory, brightness data can be flexibly determined according to the specific characteristics of the accessory, improving the accuracy of brightness extraction. By pre-calibrating the corresponding mapping relationship for different accessory types, exposure control based on the accessory region reference can also achieve personalized exposure target settings, avoiding the decrease in exposure control accuracy caused by the replacement of the reference.

[0061] In another embodiment, step S104, "controlling the exposure parameters of the image acquisition device using the target exposure brightness," includes: Based on the difference between the target exposure brightness and the current brightness of the target face, adjust at least one of the following: exposure time, analog gain, digital gain, and aperture of the image acquisition device.

[0062] In this embodiment, after determining the target exposure brightness, it is also necessary to obtain the actual brightness value of the target face region in the current image, that is, the current brightness of the target face. The difference between the target exposure brightness and the current brightness of the target face is calculated. This difference reflects the direction and degree of deviation between the current exposure level and the target exposure level. For example, if the target exposure brightness is greater than the current brightness, the difference is positive, indicating insufficient exposure and requiring increased exposure; if the target exposure brightness is less than the current brightness, the difference is negative, indicating overexposure and requiring reduced exposure; if the target exposure brightness equals the current brightness, the difference is zero, indicating that the current exposure has reached the target level and no adjustment is needed. Based on this difference value, a corresponding exposure parameter adjustment signal can be generated to adjust the exposure parameters of the image acquisition device. Adjustable exposure parameters of an image acquisition device may include at least one of the following: exposure time, which is the length of time for photoelectric pixel arrays in the image sensor to accumulate photocharge; increasing the exposure time will brighten the overall image, while decreasing the exposure time will darken the overall image; analog gain, which is the amplification factor when the image sensor converts the charge signal accumulated by the photoelectric pixels into a voltage signal; increasing the analog gain will amplify the signal amplitude, thus brightening the image, while decreasing the analog gain will darken the image; digital gain, which is the factor by which the digital signal value is multiplicatively amplified after analog-to-digital conversion; increasing the digital gain will brighten the image, while decreasing the digital gain will darken the image; aperture, which is an adjustable mechanism that controls the size of the lens's light-gathering aperture; opening the aperture increases the amount of light passing through, thus brightening the image, while closing the aperture decreases the amount of light passing through, thus darkening the image. In one example, one or more combinations of the above exposure parameters can be selected for adjustment based on the magnitude and direction of the difference value. For example, when the difference value is large, the exposure time and analog gain can be adjusted simultaneously to achieve a fast response; when the difference value is small, only the digital gain can be adjusted to achieve fine adjustment. When adjusting exposure parameters, it is also necessary to consider the mutual constraints between various parameters. For example, excessive exposure time may cause motion blur, and excessive aperture may cause shallow depth of field. Therefore, in actual control, a preset priority strategy or optimization algorithm can be used to allocate the adjustment amount of each parameter in order to achieve the best exposure control effect.

[0063] In this disclosure, at least one of the following parameters of the image acquisition device—exposure time, analog gain, digital gain, and aperture—is adjusted based on the difference between the target exposure brightness and the current brightness of the target face. Thus, by comparing the difference between the target exposure brightness and the actual current brightness to drive the adjustment of exposure parameters, the exposure control has an automatic correction capability to continuously track and maintain the target brightness level. Simultaneously, the combined use of multiple adjustable parameters such as exposure time, analog gain, digital gain, and aperture provides a rich set of adjustment methods for exposure control. The adjustment method can be flexibly selected according to different application scenarios. Furthermore, by using the difference between the target exposure brightness and the current brightness as the adjustment basis, the adjustment amount of the exposure parameters matches the actual deviation, achieving stable and accurate exposure control and avoiding brightness oscillations caused by over-adjustment.

[0064] In another embodiment, an exposure control method further includes: In response to the fulfillment of the target triggering condition, detection of the target reference area and the facial target area is initiated to determine the target exposure brightness based on the relative relationship.

[0065] In this embodiment, the target trigger condition is used to determine when to activate the exposure control method of this disclosure. When the target trigger condition is detected to be met, in response to the met event, the detection operation of the target reference region and the facial target region is initiated, that is, the exposure control method of this disclosure is activated, and the detection of the target reference region and the facial target region begins to determine the target exposure brightness based on the relative relationship, and the exposure parameters are controlled based on the target exposure brightness. If the target trigger condition is not met, the exposure control method of this disclosure is not activated, and the image acquisition device operates according to the conventional exposure control method, such as adjusting the overall brightness of the image to a preset intermediate grayscale value according to a standard automatic exposure algorithm.

[0066] In another embodiment, the target triggering condition includes at least one of the following: The target button is triggered, and the target button is used to trigger the opening and closing of the exposure control logic of the image acquisition device.

[0067] In this embodiment, the target trigger condition can be: detecting that a target button has been triggered, such as the target button being in an "on" state. The target button is a physical or virtual button that can be operated by the user, specifically used to control the on / off state of the exposure control logic of the image acquisition device. This target button can be located on the physical casing of the electronic device, for example, as a function button in a camera application; or it can be located in the user interface of a touchscreen, for example, as a clickable icon in a camera application interface. When the user operates the target button to turn it on, it indicates that the user wishes to enable a personalized exposure control method based on the relative relationship between the target reference area and the facial target area. At this time, the target trigger condition is met, and the area detection and brightness analysis process is initiated. When the user operates the target button again to turn it off, it indicates that the user wishes to deactivate the personalized exposure control method and restore the normal exposure control mode. Through the target button method, the user can manually control the activation and deactivation of the personalized exposure control function. For example, in certain special shooting scenarios, when the user believes that the normal exposure mode is better, they can actively turn off the function, while in scenarios requiring precise facial exposure, such as shooting portraits, they can actively turn on the function. This gives the user full choice and control, improving the user experience.

[0068] The target instruction is received, which is used to instruct exposure control.

[0069] In this embodiment, the target triggering condition can be: receiving a target instruction. The target instruction indicates that the personalized exposure control of this disclosure is required. The source of the target instruction can be varied. For example, when a user opens a camera application and switches to portrait shooting mode, the camera application can send a target instruction to the image processing module to instruct the activation of personalized exposure control; when the electronic device detects a face in the current shooting scene, the intelligent scene recognition module can automatically send a target instruction; when the electronic device is in video conferencing mode and detects a participant in the frame, the video conferencing application can send a target instruction. Upon receiving the target instruction, the electronic device immediately responds to the instruction and initiates the detection process for the target reference area and the facial target area. Through the target instruction method, the software programmable control of the exposure control method is realized, allowing other applications or system services to flexibly call the exposure control function according to their own logic.

[0070] The expected exposure value of the image acquisition device is determined to be greater than the first threshold.

[0071] In this embodiment, the target triggering condition can be: determining that the expected exposure value of the image acquisition device is greater than a first threshold. The expected exposure value refers to the overall brightness level or exposure of the image corresponding to the exposure parameters calculated using a conventional exposure algorithm under the current ambient lighting conditions. When the expected exposure value exceeds the first threshold, it indicates that the ambient lighting conditions may be poor or the subject itself has low reflectivity. In this case, the conventional exposure algorithm may produce exposure deviations because it cannot accurately determine whether the lighting is insufficient or the subject itself is too dark. Therefore, it is necessary to activate a personalized exposure control method based on the relative relationship between the target reference area and the facial target area to accurately determine the current situation and provide a reasonable exposure scheme. In one example, the automatic exposure module of the image acquisition device continuously monitors the current ambient lighting level and the expected exposure parameters, and calculates the corresponding expected exposure value. When the expected exposure value is greater than the preset first threshold, the triggering condition is activated, and the detection of the target reference area and the facial target area begins, improving the intelligence level and response speed of the exposure control.

[0072] In this disclosure, a target trigger condition is introduced. Detection of the target reference area and the facial target area is only initiated when the target trigger condition is met to determine the target exposure brightness. Thus, by setting the trigger condition, the continuous operation of steps such as area detection and brightness analysis is avoided, effectively reducing the system's computing resource consumption and power consumption, and extending the battery life of electronic devices. At the same time, by setting target button trigger conditions, target command trigger conditions, and automatic trigger conditions when the expected exposure value is greater than a first threshold, the manual control of personalized exposure control functions, the software programmable control of the exposure control method, and the automatic activation of the system when exposure deviations may occur are respectively realized. This makes the exposure control method take into account user choice, functional expandability, and intelligent response capabilities, and is suitable for a variety of different usage scenarios.

[0073] In another embodiment, "determining the visible sclera region in the target face" in step S202 includes: The first-resolution image containing the target face, output by the acquired image sensor, is downsized to obtain a second-resolution image; The eye position of the target face is detected based on the second resolution image, so as to crop out the eye image corresponding to the eye position from the first resolution image; The visible area of ​​the sclera is determined based on eye images.

[0074] In this embodiment, the image sensor of the image acquisition device outputs a first-resolution image containing the target face. The first-resolution image refers to image data with a specific resolution output by the image sensor according to the current operating mode; this first resolution is the full resolution of the image sensor, i.e., the maximum resolution. Regardless of the resolution of the image requested by the backend application, the image sensor outputs the first-resolution image. The first-resolution image contains complete image information of the target face, with sufficient pixels and clear details, which is beneficial for accurate scleral region segmentation. However, the first-resolution image has a large data volume, resulting in high computational load and slow processing speed when directly used for computer vision tasks such as face detection and eye localization. Therefore, it is necessary to perform a downsampling operation on the acquired first-resolution image. Downsampling refers to reducing the spatial resolution of the image, reducing the number of pixels in the image, which can be achieved through image downsampling methods. Downsampling methods can employ algorithms such as nearest neighbor interpolation, bilinear interpolation, or bicubic interpolation to merge multiple adjacent pixels in the first-resolution image into one pixel, thereby generating a lower-resolution image, i.e., a second-resolution image. The data volume of the second-resolution image after downscaling is significantly reduced. Subsequent computational operations such as face detection and eye localization can be performed on this second-resolution image, which can greatly reduce computational complexity and processing time, and improve processing efficiency.

[0075] In this embodiment, after obtaining the second-resolution image, face detection and eye localization operations are performed on the second-resolution image. Since the second-resolution image has a smaller data volume, the face detection and eye localization algorithms run quickly, enabling rapid detection results. First, the face detection algorithm is used to locate the target face in the second-resolution image, obtaining the bounding box coordinates of the target face in the second-resolution image. Then, within this bounding box, the face keypoint detection algorithm is used to locate the eye keypoints, obtaining their coordinates in the second-resolution image. After obtaining the eye coordinates in the second-resolution image, these coordinates need to be mapped back to the coordinate system of the first-resolution image. Since the size ratio between the first-resolution and second-resolution images is known, the eye position coordinates in the second-resolution image can be converted to their corresponding coordinates in the first-resolution image based on this ratio. Then, based on the converted coordinates, the region containing the eye position is determined in the first-resolution image, and a local image within this region is cropped from the first-resolution image; this local image is the eye image corresponding to the eye position. Because the first-resolution image has a high resolution, the cropped eye image retains rich details of the eye, including clear boundaries and texture features of the sclera, iris, and pupil, providing a high-quality image data foundation for subsequent accurate segmentation of the visible area of ​​the sclera.

[0076] In this embodiment, after cropping the eye image from the first-resolution image, a precise segmentation operation of the visible scleral region is performed on the eye image. Since the eye image is cropped from the high-resolution first-resolution image, the number of pixels in the eye region is sufficient and the details are clear, which is beneficial for accurate scleral region segmentation. Scleral pixels can be extracted from the eye image using color space analysis and threshold segmentation methods, as mentioned earlier and will not be repeated here. After determining the visible scleral region, the boundary of this region can be further refined, for example, by using an active contour model or graph cut algorithm, to make the boundary of the visible scleral region more accurately match the actual scleral edge. The finally determined visible scleral region is the set of scleral pixels in the eye image.

[0077] In this disclosure, a first-resolution image is downscaled to obtain a second-resolution image. The eye position is detected based on the second-resolution image, and an eye image is cropped from the first-resolution image. Then, the visible sclera region is determined based on the eye image. This allows for face detection and eye localization on low-resolution images, with small data volume and fast processing speed, significantly reducing computational complexity. After obtaining the eye position, an eye image is cropped from the original high-resolution image for sclera segmentation, preserving sufficient image details and ensuring segmentation accuracy. This scheme balances processing speed and accuracy, avoiding both the enormous computational overhead of direct detection on the high-resolution full image and the insufficient accuracy problem caused by direct segmentation on low-resolution images.

[0078] Figure 4 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 4 ; Figure 5 A schematic diagram of a scenario for an exposure control method according to an embodiment of this disclosure is shown, such as... Figure 4 As shown, the image sensor of the image acquisition device outputs raw data at maximum resolution. This raw data represents the highest resolution image data that the image sensor can provide within its own pixel scale. This raw data contains complete image information, providing the richest detail foundation for subsequent image processing. For example... Figure 5 As shown, the left side is an eye image at 480P resolution, and the right side is an eye image at 2160P resolution. It is obvious that the high-resolution eye image on the right is clearer and retains more eye details. Therefore, the image sensor needs to output raw data at the maximum resolution to improve the accuracy of the visible area of ​​the sclera in subsequent processing.

[0079] In this embodiment, the maximum resolution raw data first enters the DSP (Digital Signal Processor) data splitter interface. This splitter interface is a key node in the internal data path, its function being to simultaneously or time-divisionally distribute the input maximum resolution raw data to two independent processing branches: the classic main path and the eye-specific branch. Through the splitter interface, one data path is sent to the classic main path for conventional image processing, while the other data path is sent to the eye-specific branch for specialized analysis and processing of the eye region. The two branches work in parallel without interference, ensuring that the conventional image processing and the specialized eye analysis can be performed simultaneously.

[0080] In this embodiment, in the classic main path, the raw data at maximum resolution is processed by various image processing modules before outputting a processed image. The image processing modules included in the classic main path may include, but are not limited to, de-mosaic modules, noise reduction modules, color correction modules, automatic exposure statistics modules, and automatic white balance statistics modules. These modules perform conventional pixel-level processing operations on the raw image data, ultimately outputting a normal image with image quality restoration and color restoration for use by backend applications. These image processing modules continuously process the input raw data to ensure that the image quality output by the main path meets basic imaging requirements.

[0081] In this embodiment, the DSP typically performs image recognition based on its built-in Computer Vision (CV) unit. However, the CV unit usually cannot accept full-resolution images and has certain capacity and computing power limitations. Therefore, it is necessary to first downscale the image at the highest resolution (the first resolution image) to obtain a second resolution image supported by the CV unit. Generally, the controller at the front end of the CV unit calls the scaling down unit to proportionally downscale the image at the highest resolution to the second resolution image supported by the CV unit. Then, the CV unit performs face and eye position recognition on the second resolution image. For example, using face detection and eye localization methods, the eye position of the target face is located, and then the coordinates are sent to the controller at the front end of the CV unit. The controller at the front end of the CV unit can call the crop unit to accurately crop the eye image from the first resolution image based on the position coordinates. Of course, the correspondence between the position coordinates of the first and second resolution images has been predetermined. The controller at the front end of the CV unit can then return the cropped eye and face images from the first resolution image to the CV unit for scleral and skin brightness and color recognition.

[0082] In this embodiment, in the eye branch, the CV unit can perform eye-specific image processing on the extracted eye image. For example, it can perform enhanced brightness equalization processing to highlight the brightness difference between the sclera region and surrounding tissues; it can perform high-precision edge enhancement processing to make the boundary between the sclera, iris, and eyelid clearer, facilitating subsequent segmentation and recognition; and it can perform specialized color space conversion to convert the eye image to a color space more suitable for extracting the white features of the sclera. After processing the eye image, the visible sclera region and the facial target region are identified. For example, in the processed eye image, the location and range of the visible sclera region are located using methods such as color thresholding and connected component analysis. Simultaneously, combined with the facial key point detection results, the eye region, mouth region, hair region, and accessory region are excluded from the overall facial region of the target face to obtain the location and range of the facial target region. The brightness values ​​of each pixel in the visible sclera region and the facial target region are extracted respectively, and the first average brightness of the visible sclera region and the second average brightness of the facial target region are calculated. Finally, the second average brightness of the target facial area is divided by the first average brightness of the visible sclera area to obtain the ratio between the two. This ratio is the facial reflection characteristics of the target face, i.e., the relative relationship, and this relative relationship is sent to the classical main path.

[0083] In this embodiment, the classic main path includes an ISP (Image Signal Processor). After receiving the relative relationship output from the eye branch, the ISP determines the target exposure brightness from the target mapping table based on this relationship, and then uses this target exposure brightness to control the exposure parameters of the image acquisition device. Simultaneously, the image output from the classic main path can serve as feedback information to verify whether the imaging effect after adjusting the exposure parameters meets expectations. The image output from the classic main path and the relative relationship output from the eye branch work together to form a complete personalized exposure control closed-loop system.

[0084] In another embodiment, step S203, "determining the target exposure brightness corresponding to the target face based on the relative relationship between the brightness data of the visible sclera area and the facial target area," includes: Calculate the effective pixel ratio and brightness distribution characteristics of the visible area of ​​the sclera; The reliability of facial reflection characteristics is determined based on the effective pixel ratio and brightness distribution features; Facial reflectance characteristics are adjusted based on confidence level to determine target exposure brightness based on the adjusted facial reflectance characteristics.

[0085] In this embodiment, after determining the visible area of ​​the sclera, it is necessary to determine the effective pixel ratio and brightness distribution characteristics of the visible area. The effective pixel ratio refers to the proportion of pixels in the visible area of ​​the sclera that can be effectively used for brightness statistics to the total number of pixels in that area. The criteria for determining effective pixels can be: the pixel's brightness value is within the normal brightness range, i.e., it is neither an excessively dark pixel nor an excessively bright pixel with a brightness value close to saturation; the pixel's color characteristics conform to the white characteristics of the sclera, i.e., it is within the typical chromaticity range of the sclera in the color space. If a pixel's brightness value is abnormal (too dark or too bright) or its color characteristics deviate from the typical characteristics of the sclera, it is marked as an invalid pixel. Brightness distribution characteristics refer to the statistical distribution of the brightness values ​​of each pixel within the visible area of ​​the sclera. In one example, the variance or standard deviation of the pixel brightness values ​​within the visible area of ​​the sclera can be calculated. A larger variance or standard deviation indicates that the brightness values ​​of each pixel within the area are more dispersed and have poorer uniformity; a smaller variance or standard deviation indicates that the brightness values ​​of each pixel within the area are more concentrated and have better uniformity.

[0086] In this embodiment, after calculating the effective pixel ratio and brightness distribution characteristics of the visible sclera region, the reliability of facial reflectivity is determined based on these two indicators. Reliability represents the degree of reliability of the currently calculated facial reflectivity values. The reliability can be determined as follows: a higher effective pixel ratio indicates more effective pixels available for brightness statistics in the visible sclera region, resulting in stronger representativeness of the brightness data and higher reliability; conversely, a lower effective pixel ratio indicates a large number of pixels in the visible sclera region are unusable due to occlusion, reflection, or abnormalities, leading to poorer representativeness of the brightness data and lower reliability. A more concentrated brightness distribution characteristic (e.g., smaller variance) indicates more uniform brightness values ​​within the visible sclera region, with no abnormally bright or dark areas, resulting in higher reliability; conversely, a more dispersed brightness distribution (e.g., larger variance) indicates uneven brightness values ​​within the visible sclera region, potentially with localized reflections or shadows, resulting in lower reliability. In one example, the effective pixel ratio and brightness distribution characteristics (e.g., variance) can be compared with their respective thresholds, and the reliability level can be determined based on the comparison results. For example, the first level (high confidence) is set as follows: the effective pixel ratio is greater than 90% and the variance is less than the first variance threshold; the second level (medium confidence) is set as follows: the effective pixel ratio is between 70% and 90% and the variance is less than the second variance threshold; the third level (low confidence) is set as follows: the effective pixel ratio is less than 70% or the variance is greater than or equal to the second variance threshold.

[0087] In this embodiment, the facial reflection characteristics also need to be adjusted according to the determined confidence level. Then, the target exposure brightness is determined from the target mapping table based on the adjusted facial reflection characteristics. The adjustment operation can be as follows: When the confidence level is high (e.g., high effective pixel ratio and concentrated brightness distribution), it indicates that the imaging quality of the visible sclera area is good, and the calculated facial reflection characteristics are accurate and reliable. In this case, there is no need to adjust the facial reflection characteristics, and the original calculated value is directly used as the final facial reflection characteristics. When the confidence level is medium (e.g., medium effective pixel ratio or a certain degree of dispersion in brightness distribution), it indicates that the imaging quality of the visible sclera area is somewhat abnormal, and the calculated facial reflection characteristics may have deviations. In this case, the facial reflection characteristics can be smoothed, for example, by weighting the current calculated value with the historical facial reflection characteristic value corresponding to the previous frame image, and using the stability of historical data to suppress the influence of the current outlier. When the reliability is low (e.g., low effective pixel ratio or severely dispersed brightness distribution), it indicates that there is a serious abnormality in the imaging quality of the visible area of ​​the sclera, and the calculated facial reflection characteristics are unreliable. In this case, one of the following measures can be taken: set the facial reflection characteristics to a preset default value (e.g., the standard value corresponding to medium reflection characteristics) to avoid extreme erroneous exposure caused by abnormal data; or directly abandon the personalized exposure control based on relative relationship and revert to the regular automatic exposure mode, controlling the exposure parameters according to the standard intermediate gray value.

[0088] In this disclosure, the reliability of facial reflection characteristics is determined by calculating the effective pixel ratio and brightness distribution characteristics of the visible area of ​​the sclera, and the facial reflection characteristics are adjusted based on the reliability. This effectively assesses the imaging quality of the visible area of ​​the sclera, identifies brightness data anomalies caused by factors such as occlusion, reflection, and eye abnormalities. By differentiating facial reflection characteristics based on reliability, high reliability is used directly, medium reliability is smoothed, and low reliability is either set to a default value or reverted to the normal mode, effectively preventing exposure control errors caused by abnormal data. This quality assessment and adjustment mechanism significantly improves the robustness and stability of the exposure control method, enabling it to work reliably in various complex real-world shooting scenarios and avoiding exposure jumps or incorrect exposures caused by single-frame abnormal data.

[0089] In another embodiment, an exposure control method further includes at least one of the following: The facial reflection characteristics of the current image containing the target face are smoothed based on the facial reflection characteristics corresponding to historical images, and the target exposure brightness is determined based on the smoothed facial reflection characteristics.

[0090] In this embodiment, in a video stream or continuous preview scenario, the image acquisition device continuously outputs multiple frames of images, and each frame can yield a corresponding facial reflectance value. However, due to factors such as brightness fluctuations in video conferencing, subject movement, image noise, detection errors, or instantaneous fluctuations in ambient light, the facial reflectance values ​​calculated between adjacent frames may exhibit some degree of jitter. If the facial reflectance values ​​calculated independently for each frame are directly used to determine the target exposure brightness, it may cause frequent fluctuations in exposure parameters, resulting in flickering or instability in the image. Therefore, the facial reflectance characteristics corresponding to historical images can be used to smooth the facial reflectance characteristics corresponding to the current image. For example, the facial reflectance value corresponding to the previous frame or several previous frames can be stored in the storage area. After calculating the facial reflectance value of the current frame, an exponentially weighted moving average or a sliding window average method can be used to smooth the current value. After smoothing, the target exposure brightness is determined from the target mapping table based on the smoothed facial reflectance characteristics.

[0091] The target exposure brightness is determined by combining the facial reflection characteristics and the ambient brightness of the space in which the image acquisition device is located.

[0092] In this embodiment, the target exposure brightness can also be determined based on facial reflection characteristics and the ambient brightness of the current spatial environment in which the image acquisition device is located. Facial reflection characteristics are determined by the ratio of the brightness of the facial target area to the visible area of ​​the sclera, which is a normalized parameter independent of external light intensity. However, in actual exposure control, the target exposure brightness should consider both the facial reflection characteristics of the subject and the ambient light. For example, a dark subject in an extremely dark environment and a dark subject in a bright environment, although having the same facial reflection characteristics, require different absolute exposure parameters. Therefore, ambient light must also be considered when determining the target exposure brightness. In one example, facial reflection characteristics can be used as the primary factor, and ambient brightness as the adjustment factor. First, a basic target exposure brightness is determined from the target mapping table based on the facial reflection characteristics. Then, the target exposure brightness is corrected according to the ambient brightness; for example, the target exposure brightness is appropriately increased when the ambient brightness is low, and appropriately decreased when the ambient brightness is high. In one example, facial reflectivity and ambient brightness can be used as two independent input variables, fed into a two-dimensional mapping table. This table pre-stores target exposure brightness values ​​corresponding to different combinations of facial reflectivity and ambient brightness. By comprehensively considering both factors, personalized exposure based on the subject's facial reflectivity can be achieved while adapting the exposure parameters to the current ambient absolute illumination level. In another example, different influence weights can be assigned to facial reflectivity under different ambient brightness levels. For instance, when the ambient brightness is below a preset threshold, the influence weight of facial reflectivity on target exposure brightness is increased; when the ambient brightness is above the preset threshold, the influence weight of facial reflectivity on target exposure brightness is decreased.

[0093] The exposure compensation amount for the target face is determined based on the facial reflection characteristics, and the target exposure brightness is adjusted based on the exposure compensation amount.

[0094] In this embodiment, exposure compensation refers to the additional amount of exposure that needs to be added or reduced based on the standard exposure brightness determined by the conventional automatic exposure algorithm. In conventional automatic exposure mode, the image acquisition device usually adjusts the overall brightness of the image or the brightness of the main area of ​​the image to a preset standard intermediate gray value. The target exposure brightness corresponding to this standard intermediate gray value is recorded as the standard exposure brightness. In this embodiment, an exposure compensation amount can be determined based on the facial reflection characteristics of the current target face, and then the standard exposure brightness can be adjusted using this exposure compensation amount to obtain the final target exposure brightness. For example, when the ratio corresponding to the facial reflection characteristics is small (indicating low facial reflectivity), a positive exposure compensation amount is determined, that is, the standard exposure brightness is increased by a certain amount, so that the target exposure brightness is higher than the standard exposure brightness, to avoid the facial area being too dark. When the ratio corresponding to the facial reflection characteristics is large (indicating high facial reflectivity), a negative exposure compensation amount is determined, that is, the standard exposure brightness is decreased by a certain amount, so that the target exposure brightness is lower than the standard exposure brightness, to avoid the facial area being too bright. The exposure compensation amount can be calculated based on the facial reflection characteristic value through a linear function or a nonlinear function, or it can be obtained by looking up a pre-stored compensation mapping table.

[0095] This disclosure provides optimization schemes such as smoothing facial reflection characteristics based on historical images, determining target exposure brightness based on facial reflection characteristics and ambient brightness, and determining exposure compensation based on facial reflection characteristics. Thus, smoothing effectively suppresses inter-frame jitter, making changes in facial reflection characteristics smoother and avoiding image flickering caused by frequent fluctuations in exposure parameters; determining target exposure brightness in conjunction with ambient brightness allows for personalized exposure while adapting to the absolute ambient light level, making exposure control more comprehensive and accurate; adjusting target exposure brightness by determining exposure compensation can be directly extended to existing exposure control system architectures, offering good compatibility and ease of deployment.

[0096] In another embodiment, step S305, "determining the target exposure brightness for the target face from the target mapping table based on facial reflection characteristics," includes at least one of the following: In response to the target face's proportion in the image being less than a second threshold, the target exposure brightness is determined from the target mapping table based on facial reflectivity and ambient brightness; the weight of ambient brightness is greater than the weight of facial reflectivity.

[0097] In this embodiment, when the proportion of the target face in the image is less than the second threshold, it indicates that the face area in the image is relatively small compared to the overall image area, for example, the target person is far from the image acquisition device. In this case, the brightness change of the face area has a relatively limited impact on the overall visual perception of the image, while the brightness level of the background area has a more significant impact on the overall image quality. In this situation, when determining the target exposure brightness, both facial reflectivity and ambient brightness should be considered simultaneously, with the weight of ambient brightness being greater than that of facial reflectivity. For example, when determining the target exposure brightness, a first exposure brightness value is determined from the target mapping table based on facial reflectivity, and a second exposure brightness value is determined from the ambient brightness mapping table based on ambient brightness. Then, the first and second exposure brightness values ​​are weighted and summed, with the weight of the second exposure brightness value being greater than that of the first exposure brightness value, thus obtaining the final target exposure brightness. For example, the weight of ambient brightness can be set to 70%, and the weight of facial reflectivity to 30%.

[0098] In response to the presence of multiple target faces in an image, and the proportion of each target face in the image being greater than a third threshold, the target exposure brightness is determined from the target mapping table based on the average value of the facial reflection characteristics of the multiple target faces.

[0099] In this embodiment, when multiple target faces exist in an image, and the proportion of each of these target faces in the image is greater than the third threshold, it indicates that multiple faces simultaneously occupy a large proportion of the image, and each target face has a significant impact on the overall image exposure. In this case, the target exposure brightness cannot be determined solely based on the facial reflection characteristics of one face; instead, the facial reflection characteristics of all target faces should be considered comprehensively. For example, the scleral visible area and facial target area of ​​each target face can be determined separately, and the facial reflection characteristic value corresponding to each target face can be calculated separately. Then, the facial reflection characteristic values ​​of all target faces can be arithmetically averaged to obtain the average facial reflection characteristic. The target exposure brightness can then be determined from the target mapping relationship table based on this average facial reflection characteristic. In another embodiment, a weighted average method can be used, assigning different weights according to the proportion of each target face, with faces having a larger proportion receiving a higher weight, and then calculating the weighted average. This allows the determined target exposure brightness to take into account the exposure effect of all faces in the image to the greatest extent, avoiding overexposure or underexposure of other faces due to bias towards one face.

[0100] In response to the presence of multiple target faces in an image, and at least one target face having a proportion of less than a fourth threshold in the image, the target exposure brightness is determined from the target mapping table based on the average value of the facial reflectance characteristics of the other target faces besides at least one target face.

[0101] In this embodiment, when multiple target faces exist in an image, and at least one of these target faces occupies a proportion of less than a fourth threshold, it indicates that both primary and secondary faces exist in the image. The primary face occupies a larger area, while the secondary face occupies only a small proportion. In this case, the secondary face has a very limited impact on the overall image exposure due to its small size. Including it in the comprehensive calculation of facial reflection characteristics may introduce unnecessary errors due to insufficient representativeness of its brightness data. Therefore, when performing facial reflection characteristic analysis, secondary target faces with a proportion less than the fourth threshold should be excluded, and the exposure target should be determined only based on the facial reflection characteristics of the remaining target faces with a proportion greater than or equal to the fourth threshold. For example, firstly, the proportion of each target face in the image is calculated, and the proportion is compared with the fourth threshold. Target faces with a proportion greater than or equal to the fourth threshold are selected as valid faces. Then, the average facial reflection characteristics of all valid faces are calculated. Finally, the target exposure brightness is determined from the target mapping table based on this average value. This avoids the interference of brightness data from minor, small-area faces on exposure decisions, ensuring that exposure control primarily serves the main subject in the image.

[0102] In this disclosure, when the target face occupies a small proportion, the ambient brightness weight is higher than the facial reflection characteristic weight, so that the exposure control takes into account the overall image quality and avoids sacrificing background exposure due to excessive focus on small face areas; when multiple faces occupy a large proportion, the exposure target is determined by comprehensively considering the facial reflection characteristics of all faces, so that the exposure effect can take into account all faces in the image; when there are small secondary faces, they are excluded and the exposure target is determined only based on the main faces, avoiding interference from the brightness data of small areas on the exposure decision and improving the accuracy of exposure control.

[0103] In another embodiment, step S305, "determining the target exposure brightness for the target face from the target mapping table based on facial reflection characteristics," includes: In response to the facial reflectivity being less than the first reference value, the standard exposure brightness in the target mapping table is reduced by the first exposure compensation amount to obtain the target exposure brightness; In response to the facial reflectivity being less than the second reference value, it is determined that the difference between the target exposure brightness and the standard exposure brightness in the target mapping table is less than the fifth threshold; the second reference value is greater than the first reference value. In response to the facial reflectivity being greater than the second reference value, the standard exposure brightness in the target mapping table is increased by the second exposure compensation amount to obtain the target exposure brightness.

[0104] In this embodiment, the first reference value is a preset threshold used to define whether the facial reflectivity is at a low level. When the facial reflectivity is less than the first reference value, it indicates that the facial reflectivity of the target face is low, meaning the brightness of the facial target area is much lower than the brightness of the visible sclera area. In this case, if the exposure is still performed according to the standard exposure brightness (i.e., the general brightness target set by the traditional automatic exposure algorithm for objects with medium reflectivity), the facial area will be overexposed, potentially causing overexposure of the background and bright objects, and loss of detail. Therefore, in this embodiment, the standard exposure brightness is reduced by the first exposure compensation amount, that is, the exposure target is shifted towards a darker direction to obtain the final target exposure brightness. By reducing the target exposure brightness, facial area distortion and background detail loss caused by overexposure can be avoided, making the brightness of the target face in the final captured image more consistent with its low reflectivity. The first exposure compensation amount can be related to the degree to which the facial reflectivity deviates from the first reference value; the greater the deviation, the greater the first exposure compensation amount, to achieve more precise differential adjustment.

[0105] In this embodiment, the second reference value is a preset threshold value, which is greater than the first reference value, and is used to define whether the facial reflectivity belongs to the upper boundary of a medium level. When the facial reflectivity is less than the second reference value but greater than or equal to the first reference value, it indicates that the facial reflectivity of the target face is at a medium level, that is, the brightness of the facial target area is within a medium proportion of the brightness of the visible area of ​​the sclera. In this case, the facial reflectivity of the target face is close to the standard object reflectivity assumed by traditional automatic exposure algorithms, so there is no need to make a large adjustment to the standard exposure brightness. For example, the difference between the target exposure brightness and the standard exposure brightness can be set to be less than the fifth threshold, that is, the target exposure brightness is limited to a small range near the standard exposure brightness. The target exposure brightness can be set to be equal to the standard exposure brightness, or it can be set to be a small adjustment less than the fifth threshold based on the standard exposure brightness.

[0106] In this embodiment, when the facial reflectance is greater than the second reference value, it indicates that the target face has a high facial reflectance, meaning the brightness of the target facial area is close to or even reaches the brightness level of the visible sclera. In this case, if the standard exposure brightness is still applied, the facial area will be excessively underexposed, potentially resulting in a dark image, underexposed background, and a poor-looking appearance of the target face. Therefore, the standard exposure brightness can be increased by the second exposure compensation, shifting the exposure target towards a brighter direction to obtain the final target exposure brightness. By increasing the target exposure brightness, the dark facial area and loss of background details caused by underexposure can be avoided, making the brightness of the target face in the final captured image more consistent with its high reflectance characteristics. The second exposure compensation can be related to the degree to which the facial reflectance deviates from the second reference value; the greater the deviation, the greater the second exposure compensation, to achieve more precise differential adjustment.

[0107] In this disclosure, the exposure brightness of low-reflectivity targets (faces) is reduced to avoid overexposure and loss of detail caused by excessive brightness increase; the exposure brightness of medium-reflectivity targets (faces) is kept close to the standard exposure brightness to ensure stable exposure control under normal conditions; and the exposure brightness of high-reflectivity targets (faces) is increased to avoid dark images and loss of detail caused by underexposure. This achieves precise, personalized exposure control.

[0108] Figure 6 A flowchart illustrating an exposure control method according to an embodiment of this disclosure is shown. Figure 5 ,like Figure 6 As shown, a complete embodiment of an exposure control method in this disclosure may be: The camera continuously acquires image data and outputs raw images. After the raw image data flows into the processing chain, the system performs face detection and facial landmark localization on the raw image. After completing the face detection and landmark localization operations, the system determines whether a face has been detected. If the result is that no face has been detected, the system directly jumps to run the traditional AE (Auto Exposure) algorithm with global automatic exposure. It relies on the traditional global exposure logic to complete the image exposure adjustment processing, and finally outputs a final image with a natural face and preserved background details. If the result is that a face has been successfully detected, the image data enters the following processing flow: steps 1-4. Step 1: Precise ROI Extraction. This process uses the identified face region as the basic processing area. Two parallel processing branches simultaneously perform region segmentation. One processing branch first locates the eye socket and eyeball areas based on the parsed key eye points, and then segments the sclera region based on the locked eye socket and eyeball areas, marking this sclera region as ROI_E. The other processing branch first removes the eye area, eyebrow area, lip area, and hair area within the defined face region, and then extracts the clean facial target region from the face region after removing interference areas, marking this facial target region as ROI_S.

[0109] Step 2: Calculate the brightness to contrast ratio. The system calculates the average brightness value L_sclera for the ROI_E region and the average brightness value L_SK for the ROI_S region. After obtaining the average brightness parameters for the two regions, the system calculates the key ratio according to the formula R=L_SK / L_sclera. The key ratio obtained is the facial reflection characteristics of the corresponding human face.

[0110] Step 3: Dynamic Mapping and Exposure Decision. The facial reflection characteristic parameter R is fed into the dynamic mapping and exposure decision process corresponding to Step 3 as an input parameter. The system retrieves the preset stored dynamic mapping table between R value and target brightness for lookup and matching calculation. After the mapping table matching and conversion, a personalized target brightness value adapted to the current face's own attributes is output.

[0111] Step 4: Perform exposure optimization. The exposure optimization process combines personalized target brightness values ​​to adjust and control the exposure parameters at the image acquisition end, outputting a final image with a natural-looking face and preserved background details.

[0112] It is important to emphasize that the exposure control method disclosed herein can also be applied to scenarios where the target face is wearing sunglasses or colored contact lenses. In scenarios where sunglasses are worn, the lenses have visible light blocking and filtering properties, completely obscuring the user's eye area. This results in ordinary image acquisition devices (such as RGB cameras) failing to capture detailed features of the sclera, leading to a lack of effective eye imaging information. This imaging defect directly causes misjudgments when subsequently determining facial reflection characteristics. Therefore, when acquiring images containing the target face, the CV AI unit of the device's DSP can simultaneously capture bright frames when the IRLED (infrared light-emitting diode) is on and crop out the original resolution image of the eyes. Relying on the IR infrared light actively emitted by the device, which has the characteristic of penetrating conventional sunglasses lenses, the obstruction limitations of visible light imaging can be effectively overcome, resulting in clear black-and-white grayscale images of the face and eyes, fully restoring eye details. The subsequent data processing is similar to that of images acquired by ordinary image acquisition devices. That is, grayscale difference analysis and data calculation are performed on the human eye area of ​​the IR bright frame image to accurately predict the facial reflectivity characteristics of the user, effectively avoid imaging interference caused by sunglasses wearing scenarios, and ensure the continuous and stable operation of the facial reflectivity detection process.

[0113] In one example, the IR path image source can be a standalone IR camera or IR path image data from a hybrid image sensor. That is, when acquiring an image containing a target face in this disclosure, it is possible to acquire only the visible light image containing the target face acquired by the image sensor, or to acquire the visible light image containing the target face while simultaneously acquiring the infrared light image containing the target face from a standalone infrared sensor or the infrared acquisition path of the image sensor.

[0114] In scenarios involving the wearing of colored contact lenses, although the sclera portion will be reduced in size depending on the type of colored contact lens, since this disclosure acquires images at the first resolution, i.e., the maximum resolution image, the reduction in the size of the sclera portion will not lead to a decrease in recognition accuracy or even prevent recognition. As long as the sclera can be detected in the target face image, the exposure control method in this disclosure can be used without affecting the control accuracy.

[0115] Figure 7 A schematic diagram of the composition structure of an electronic device according to an embodiment of the present disclosure is shown, such as... Figure 7 As shown, an electronic device includes a device body and an image acquisition device disposed on the device body, wherein the image acquisition device includes an image sensor, and the device body also includes a signal processor; Image sensors are used to: acquire images containing target human faces; The signal processor is used to: acquire an image containing a target face from an image sensor; determine the brightness data of a target reference region and a facial target region in the image, and determine the relative relationship between the brightness data of the target reference region and the facial target region; based on the relative relationship, determine the target exposure brightness corresponding to the target face, and use the target exposure brightness to control the exposure parameters of the image acquisition device.

[0116] In one example, the signal processor may include an image signal processor (ISP) and a digital signal processor (DSP); after obtaining the relative relationship between the brightness data of the target reference area and the facial target area, the DSP can send the relative relationship to the image signal processor, which will then determine the target exposure brightness and exposure parameters.

[0117] It should be noted that the description of the electronic device in this application is similar to the description of the method embodiments described above, and has similar beneficial effects as the method embodiments; therefore, it will not be repeated. Any technical details not covered in the electronic device provided in the embodiments of this application can be understood based on the description of the method embodiments described above.

[0118] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this disclosure can be achieved, and this is not limited herein.

[0119] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "a plurality of" means two or more, unless otherwise explicitly specified.

[0120] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure. Therefore, the scope of protection of this disclosure should be determined by the scope of the claims.

Claims

1. An exposure control method, comprising: Obtain an image containing the target face; Determine the brightness data of the target reference region and the facial target region in the image; Based on the relative relationship between the brightness data of the target reference area and the facial target area, the target exposure brightness corresponding to the target face is determined. The exposure parameters of the image acquisition device are controlled using the target exposure brightness.

2. The method according to claim 1, wherein the target reference region is the visible scleral region of the target face, and determining the brightness data of the target reference region and the facial target region in the image includes: Determine the brightness data of the visible scleral area and the target facial area of ​​the target face; Correspondingly, determining the target exposure brightness corresponding to the target face based on the relative relationship between the brightness data of the target reference region and the facial target region includes: Based on the relative relationship between the brightness data of the visible area of ​​the sclera and the target area of ​​the face, the target exposure brightness corresponding to the target face is determined.

3. The method according to claim 2, wherein determining the target exposure brightness corresponding to the target face based on the relative relationship between the brightness data of the visible sclera region and the facial target region includes: Determine a first average brightness of the visible area of ​​the sclera and a second average brightness of the target facial area; The facial reflection characteristics of the target face are determined based on the ratio between the second average brightness and the first average brightness. The target exposure brightness for the target face is determined from the target mapping table based on the facial reflection characteristics, wherein different facial reflection characteristics correspond to different target exposure brightness.

4. The method according to claim 2, wherein determining the brightness data of the visible scleral region and the target facial region of the target face includes: Keypoint detection is performed on the image to determine the region of interest (ROI) based on the detected eye keypoints, and the visible sclera region is determined within the ROI. The target facial region is obtained by excluding at least one of the eye region of interest, mouth region, hair region, and accessory region from the facial region of the target face. The pixel brightness values ​​in the visible area of ​​the sclera and the target area of ​​the face are filtered and / or outlier removed to obtain the corresponding brightness data.

5. The method according to claim 1, wherein the target reference region is the accessory region formed by the accessory worn by the target face, and the step of determining the target exposure brightness corresponding to the target face based on the relative relationship between the brightness data of the target reference region and the facial target data includes: The brightness data of the accessory area is determined based on the accessory information obtained from identifying the accessory area; Based on the relative relationship between the brightness data of the accessory area and the facial target area, the target exposure brightness corresponding to the target face is determined.

6. The method according to any one of claims 1 to 5, wherein controlling the exposure parameters of the image acquisition device using the target exposure brightness comprises: Based on the difference between the target exposure brightness and the current brightness of the target face, at least one of the exposure time, analog gain, digital gain and aperture of the image acquisition device is adjusted; And / or, the method further includes: In response to the fulfillment of the target triggering condition, detection of the target reference region and the facial target region is initiated to determine the target exposure brightness based on the relative relationship; The target triggering condition includes at least one of the following: The target button is triggered, and the target button is used to trigger the opening and closing of the exposure control logic of the image acquisition device; A target instruction is received, which is used to instruct exposure control; The expected exposure value of the image acquisition device is determined to be greater than a first threshold.

7. The method according to any one of claims 2 to 4, wherein, Determining the visible sclera region in the target face includes: The first resolution image containing the target face, output by the acquired image sensor, is downsized to obtain a second resolution image; The eye position of the target face is detected based on the second resolution image, so as to crop out the eye image corresponding to the eye position from the first resolution image; The visible area of ​​the sclera is determined based on the eye image.

8. The method according to claim 3, wherein determining the target exposure brightness corresponding to the target face based on the relative relationship between the brightness data of the visible sclera region and the facial target region includes: Calculate the effective pixel ratio and brightness distribution characteristics of the visible area of ​​the sclera; The reliability of the facial reflection characteristics is determined based on the effective pixel ratio and brightness distribution features. The facial reflection characteristics are adjusted based on the confidence level, and the target exposure brightness is determined based on the adjusted facial reflection characteristics; And / or, The method further includes at least one of the following: The facial reflection characteristics of the current image containing the target face are smoothed based on the facial reflection characteristics corresponding to the historical images, and the exposure brightness of the target is determined based on the smoothed facial reflection characteristics. The target exposure brightness is determined based on the facial reflection characteristics and the ambient brightness of the current spatial environment in which the image acquisition device is located. Based on the facial reflection characteristics, an exposure compensation amount is determined for the target face, and the target exposure brightness is adjusted based on the exposure compensation amount.

9. The method according to claim 3, wherein determining the target exposure brightness for the target face from the target mapping table based on the facial reflectivity includes at least one of the following: In response to the target face's proportion in the image being less than a second threshold, the target exposure brightness is determined from the target mapping table based on the facial reflection characteristics and ambient brightness; the weight of the ambient brightness is greater than the weight of the facial reflection characteristics. In response to the presence of multiple target faces in the image, and the proportion of each of the multiple target faces in the image being greater than a third threshold, the target exposure brightness is determined from the target mapping relationship table based on the average value of the facial reflection characteristics of the multiple target faces. In response to the presence of multiple target faces in the image, and at least one target face having a proportion of less than a fourth threshold in the image, the target exposure brightness is determined from the target mapping table based on the average value of the facial reflection characteristics of other target faces besides the at least one target face. In response to the facial reflectivity being less than a first reference value, the standard exposure brightness in the target mapping table is reduced by a first exposure compensation amount to obtain the target exposure brightness; or, In response to the facial reflectivity being greater than the second reference value, the standard exposure brightness in the target mapping table is increased by a second exposure compensation amount to obtain the target exposure brightness.

10. An electronic device comprising a device body and an image capturing device provided to the device body, wherein, The image acquisition device includes an image sensor, and the device body also includes a signal processor; The image sensor is used to: acquire images containing the target human face; The signal processor is configured to: acquire the image containing the target face from the image sensor; determine the brightness data of the target reference region and the facial target region in the image, and determine the relative relationship between the brightness data of the target reference region and the facial target region; Based on the relative relationship, the target exposure brightness corresponding to the target face is determined, and the exposure parameters of the image acquisition device are controlled using the target exposure brightness.