Image-specific point detection method and device, display device, electronic device, storage medium and chip

Abstract

The application relates to the field of computer vision, in particular to an image special point detection method and device, a display device, an electronic device, a storage medium and a chip, which comprises the following steps: acquiring a biological feature image, dividing the biological feature image into a background region and a foreground region; constructing a statistical data structure matched with pixel points of the biological feature image, and determining a value of the statistical data structure according to a result of the region division; when the biological feature image is updated, updating the value of the statistical data structure, wherein the value represents a frequency of each pixel point in the biological feature image being divided into the background region; and in the case that a first value of the statistical data structure reaches a first preset threshold, determining a pixel point of the biological feature image corresponding to the first value as a special point. Through the method in the present application, the special point in the biological feature image can be effectively and accurately identified.

Description

Image outlier detection methods and apparatus, display devices, electronic devices, storage media and chips Technical Field

[0001] This disclosure relates to the field of computer vision, and more particularly to a method and apparatus for detecting image outliers, a display device, an electronic device, a storage medium, and a chip. Background Technology

[0002] Capacitive sensing biometric image technology is a relatively reliable biometric identification method. It acquires image data by measuring the capacitance changes between human biometric features (such as fingerprints and palm prints) and sensors, and then performs identification and matching. This technology is widely used in various mobile smart terminals, door locks, and automobiles.

[0003] In biometric image recognition, images need to undergo preprocessing before entering the recognition process to reduce irrelevant information and improve clarity. Preprocessing steps typically include image denoising, region segmentation, and foreground enhancement. A key step is accurately distinguishing between the foreground region (the region containing biometric information, such as the fingerprint ridges in a fingerprint image) and the background region (the region not containing biometric information, such as the non-fingerprint ridge area in a fingerprint image) in order to perform targeted processing on the foreground portion containing biometric information.

[0004] However, during the manufacturing and use of capacitor modules, capacitor plate failures, short circuits in sensor circuits, and other issues can cause outliers in biometric images. It's important to note that image processing operations like erosion and dilation alone cannot remove these outliers. These outliers may be incorrectly classified as background areas instead of foreground areas during the region segmentation stage. Such misclassification directly impacts the preprocessing and recognition of biometric images, ultimately leading to recognition errors and a reduced recognition rate. Summary of the Invention

[0005] In view of this, this disclosure proposes an image outlier detection scheme.

[0006] According to one aspect of this disclosure, an image outlier detection method is provided, comprising:

[0007] A biometric image is acquired, and the biometric image is divided into background and foreground regions. A statistical data structure matching the pixels of the biometric image is constructed, and the value of the statistical data structure is determined based on the region division result. When the biometric image is updated, the value of the statistical data structure is updated, and the value represents the frequency at which each pixel in the biometric image is divided into the background region. When a first value in the statistical data structure reaches a first preset threshold, the biometric image pixel corresponding to the first value is identified as an outlier.

[0008] In one possible implementation, the division of the biometric image into background and foreground regions includes: obtaining the grayscale value of the biometric image; and for regions in the biometric image where the grayscale value is 0 or 255, classifying the regions with grayscale values ​​of 0 or 255 into the background region.

[0009] In one possible implementation, the division of the biometric image into background and foreground regions includes: for regions in the biometric image where there are no grayscale values ​​of 0 or 255, converting the grayscale values ​​of the pixels in the biometric image into corresponding signal quantities; and dividing the pixels of the biometric image into the background and foreground regions respectively based on the signal quantities.

[0010] In one possible implementation, the step of dividing the pixels of the biometric image into the background region and the foreground region according to the semaphore includes: sorting the semaphore according to a preset arrangement rule; selecting a first reference value from the sorted semaphore according to a first preset ratio; multiplying the first reference value by a second preset ratio to obtain a classification threshold; and classifying the pixels corresponding to semaphores smaller than the classification threshold into the background region, and classifying the remaining pixels into the foreground region.

[0011] In one possible implementation, constructing a statistical data structure that matches the pixels of the biometric image includes: obtaining the resolution of the biometric image; and constructing a statistical data structure with the same size as the resolution.

[0012] In one possible implementation, determining the value of the statistical data structure based on the result of region division includes: increasing the value of the statistical data structure corresponding to the background region by a first preset value, and decreasing the value of the statistical data structure corresponding to the foreground region by a second preset value, wherein the absolute value of the first preset value is less than the absolute value of the second preset value.

[0013] In one possible implementation, when the biometric image is updated, the numerical update of the statistical data structure includes: when the biometric image is updated, updating the numerical values ​​of the statistical data structure corresponding to the pixels in the background region and the foreground region according to the first preset value and the second preset value.

[0014] According to another aspect of this disclosure, an image outlier detection apparatus is provided, comprising:

[0015] Image segmentation module: used to acquire biometric images and segment the biometric images into background and foreground regions; Statistical data structure construction module: used to construct a statistical data structure that matches the pixels of the biometric images and determine the value of the statistical data structure based on the region segmentation results; Value update module: used to update the value of the statistical data structure when the biometric images are updated, the value representing the frequency at which each pixel in the biometric images is segmented into the background region; Outlier determination module: used to determine the biometric image pixel corresponding to the first value as an outlier when the statistical data structure has a first value that reaches a first preset threshold.

[0016] In one possible implementation, the image segmentation module is used to: obtain the grayscale value of the biometric image; and for regions in the biometric image where the grayscale value is 0 or 255, segment the regions in the biometric image where the grayscale value is 0 or 255 into the background region.

[0017] In one possible implementation, the image segmentation module is used to: convert the gray values ​​of the pixels in the biometric image into corresponding signal quantities for regions in the biometric image where there are no gray values ​​of 0 or 255; and divide the pixels of the biometric image into the background region and the foreground region according to the signal quantities.

[0018] In one possible implementation, the image segmentation module is used to: sort the semaphores according to a preset arrangement rule; select a first reference value from the sorted semaphores according to a first preset ratio; multiply the first reference value by a second preset ratio to obtain a classification threshold; and, according to the classification threshold, classify the pixels corresponding to the semaphores smaller than the classification threshold into the background region, and classify the remaining pixels into the foreground region.

[0019] In one possible implementation, the statistical data structure construction module is used to: obtain the resolution of the biometric image; and construct a statistical data structure with the same size as the resolution.

[0020] In one possible implementation, the statistical data structure construction module is used to: increase the value of the statistical data structure corresponding to the background region by a first preset value, and decrease the value of the statistical data structure corresponding to the foreground region by a second preset value, wherein the absolute value of the first preset value is less than the absolute value of the second preset value.

[0021] In one possible implementation, the numerical update module is used to: update the numerical values ​​of the statistical data structure corresponding to the pixels in the background region and the foreground region according to the first preset value and the second preset value when the biometric image is updated.

[0022] According to another aspect of this disclosure, a display device is provided, including a plurality of display units and at least one of the above-described image outlier detection devices.

[0023] In one possible implementation, the display unit includes a display panel, which includes at least one of a liquid crystal display panel, a micro light-emitting diode display panel, a light-emitting diode display panel, a mini light-emitting diode display panel, a quantum dot light-emitting diode display panel, an organic light-emitting diode display panel, a cathode ray tube display panel, a digital light processing display panel, a field emission display panel, a plasma display panel, an electrophoretic display panel, an electrowetting display panel, and a small-pitch display panel.

[0024] According to another aspect of this disclosure, an electronic device is provided, including the display device described above.

[0025] According to another aspect of this disclosure, an electronic device is provided, comprising:

[0026] processor;

[0027] Memory used to store processor-executable instructions;

[0028] The processor is configured to implement any of the above methods when executing instructions stored in the memory.

[0029] According to another aspect of this disclosure, a non-volatile computer-readable storage medium is provided that stores computer program instructions thereon, wherein the computer program instructions, when executed by a processor, implement the above-described method.

[0030] According to another aspect of this disclosure, a computer program product is provided, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device performs the above-described method.

[0031] According to another aspect of this disclosure, a chip is provided that includes the above-described image outlier detection device.

[0032] In this embodiment, a biometric image is acquired and divided into background and foreground regions. A statistical data structure matching the pixels of the biometric image is constructed, and the value of the statistical data structure is determined based on the region division result. When the biometric image is updated, the value of the statistical data structure is updated, whereby the value represents the frequency with which each pixel in the biometric image is classified as a background region. When a first value in the statistical data structure reaches a first preset threshold, the biometric image pixel corresponding to the first value is identified as an outlier. Thus, the characteristic changes of each pixel in multiple acquired biometric images can be dynamically tracked.

[0033] Pixels in the foreground region typically have features such as grayscale values ​​and signal intensity related to biometrics. These features change depending on the biometric pressing conditions. Therefore, in multiple images, the classification of these pixels as foreground or background regions will vary, with the frequency of being classified as background regions being relatively low and somewhat random.

[0034] Outliers are abnormal pixels caused by hardware malfunctions or other reasons. Their grayscale values ​​or signal strengths differ significantly from normal pixels, and this difference is fixed and unaffected by changes in the biometric image. Therefore, in multiple biometric images, pixels at outliers are often repeatedly and incorrectly classified as background areas, causing their values ​​in the statistical data structure to accumulate and eventually reach a first preset threshold.

[0035] When the value reaches the first preset threshold, the pixel can be accurately identified as an outlier. This enables effective and accurate identification of outliers in biometric images, reducing their impact on biometric recognition and improving the accuracy and reliability of biometric identification.

[0036] Other features and aspects of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0037] The accompanying drawings, which are included in and form part of this specification, illustrate exemplary embodiments, features, and aspects of this disclosure together with the specification and serve to explain the principles of this disclosure.

[0038] Figure 1 shows a schematic diagram of a particular point according to an embodiment of the present disclosure.

[0039] Figure 2 shows a schematic diagram of the morphology of a singularity according to an embodiment of the present disclosure.

[0040] Figure 3 shows a flowchart of an image outlier detection method according to an embodiment of the present disclosure.

[0041] Figure 4 shows a flowchart of biometric image recognition according to an embodiment of the present disclosure.

[0042] Figure 5 shows a schematic histogram of signal magnitude distribution corresponding to a biometric image according to an embodiment of the present disclosure.

[0043] Figure 6 shows a schematic histogram of signal magnitude distribution corresponding to another biometric image according to an embodiment of the present disclosure.

[0044] Figure 7 shows a schematic diagram of the result of dividing the background and foreground regions of a biometric image according to an embodiment of the present disclosure.

[0045] Figure 8 shows a schematic diagram of the updated biometric image background and foreground region segmentation according to an embodiment of the present disclosure.

[0046] Figure 9 shows a schematic diagram of fingerprint image region division and partial magnification according to an embodiment of the present disclosure.

[0047] Figure 10 shows a block diagram of an image outlier detection apparatus according to an embodiment of the present disclosure.

[0048] Figure 11 shows a block diagram of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0049] Various exemplary embodiments, features, and aspects of this disclosure will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0050] As used herein, the terms “comprising,” “including,” “having,” or variations thereof are open-ended and include one or more of the stated features, integrals, elements, steps, components, or functions, but do not exclude the presence or addition of one or more other features, integrals, elements, steps, components, functions, or groups thereof.

[0051] When an element is referred to as “connected,” “coupled,” “responding,” or a variation thereof relative to another element, it may be directly connected, coupled, or responding to another element, or there may be an intermediate element present.

[0052] Although the terms first, second, third, etc., may be used herein to describe various elements / operations, these elements / operations should not be limited by these terms. These terms are only used to distinguish one element / operation from another. Therefore, without departing from the teachings of the inventive concept, a first element / operation in some embodiments may be referred to as a second element / operation in other embodiments.

[0053] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0054] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0055] Capacitive sensing biometric image technology is a relatively reliable biometric identification method. It acquires image data by measuring the capacitance changes between human biometric features (such as fingerprints and palm prints) and sensors, and then performs identification and matching. This technology is widely used in various mobile smart terminals, door locks, and automobiles.

[0056] In biometric image recognition, images need to undergo preprocessing before entering the recognition process to reduce irrelevant information and improve clarity. Preprocessing steps typically include image denoising, region segmentation, and foreground enhancement. A key step is accurately distinguishing between the foreground region (the region containing biometric information, such as the fingerprint ridges in a fingerprint image) and the background region (the region not containing biometric information, such as the non-fingerprint ridge area in a fingerprint image) in order to perform targeted processing on the foreground portion containing biometric information.

[0057] However, during the manufacturing and use of capacitor modules, capacitor plate failures, short circuits in sensor circuits, and other issues can cause outliers in biometric images. It's important to note that image processing operations like erosion and dilation alone cannot remove these outliers. These outliers may be incorrectly classified as background areas instead of foreground areas during the region segmentation stage. Such misclassification directly impacts the preprocessing and recognition of biometric images, ultimately leading to recognition errors and a reduced recognition rate.

[0058] For example, FIG1 shows a schematic diagram of a singularity according to an embodiment of the present disclosure. Referring to FIG1, FIG1A shows a schematic diagram of the location of the singularity when a biometric image is present, and FIG1B shows a schematic diagram of the location of the singularity when a biometric image is absent. Furthermore, the shape of the singularity includes a single singularity (i.e., a pixel in the biometric image) and a connected region obtained by connecting singularities (i.e., a region formed by connecting single singularities).

[0059] For example, Figure 2 illustrates a schematic diagram of the morphology of a singularity according to an embodiment of the present disclosure. Figure A in Figure 2 illustrates a schematic diagram of the morphology of a single singularity, and Figure B in Figure 2 illustrates a schematic diagram of the morphology of a connected region obtained by connecting single singularities. It is worth noting that these singularities cannot be removed by image processing operations such as erosion and dilation alone. These singularities may be incorrectly classified into the background region during the region segmentation stage, rather than correctly identified as singularities in the foreground region. Once such incorrect segmentation occurs, it will directly affect the preprocessing and recognition process of biometric images, ultimately causing recognition errors and reducing the recognition rate.

[0060] In this embodiment, a biometric image is acquired and divided into background and foreground regions. A statistical data structure matching the pixels of the biometric image is constructed, and the value of the statistical data structure is determined based on the region division result. When the biometric image is updated, the value of the statistical data structure is updated, whereby the value represents the frequency with which each pixel in the biometric image is classified as a background region. When a first value in the statistical data structure reaches a first preset threshold, the biometric image pixel corresponding to the first value is identified as an outlier. Therefore, the characteristic changes of each pixel in multiple images can be dynamically tracked. Since outliers are not affected by fingerprint changes, pixels at outliers are often continuously and incorrectly classified as background regions in multiple acquired biometric images, causing their values ​​in the statistical data structure to accumulate and increase. When the value in the statistical data structure reaches the first preset threshold, the pixel can be accurately identified as an outlier, thus effectively and accurately identifying outliers in biometric images.

[0061] Figure 3 shows a flowchart of an image outlier detection method according to an embodiment of the present disclosure. As shown in Figure 3, the image outlier detection method according to an embodiment of the present disclosure may include:

[0062] S11. Acquire a biometric image, and divide the biometric image into a background region and a foreground region;

[0063] Biometric images can be acquired through a capacitance module, which generates biometric images by measuring the capacitance change between the biological surface and the sensor. The images can be fingerprint images, palm print images, etc., and the acquired biometric images can be grayscale images.

[0064] After acquiring the biometric image, it can be preprocessed. Figure 4 shows a flowchart of biometric image recognition according to an embodiment of the present disclosure. As shown in Figure 4, image preprocessing may include image denoising, region segmentation, and foreground region enhancement. In one possible implementation, the image outlier detection method of the present disclosure embodiment can be performed during region segmentation.

[0065] In one possible implementation, the division of the biometric image into background and foreground regions includes: obtaining the grayscale value of the biometric image; and for regions in the biometric image where the grayscale value is 0 or 255, classifying the regions with grayscale values ​​of 0 or 255 into the background region.

[0066] Grayscale values ​​describe the brightness information of each pixel in a biometric image. In digital images, they are typically represented by integers between 0 and 255, where 0 represents pure black, 255 represents pure white, and values ​​in between represent different grayscale levels. If a biometric image contains areas with grayscale values ​​of 0 or 255, these areas may be non-biological regions or regions containing anomalous features, and can be directly classified as background areas.

[0067] In this embodiment, by acquiring the grayscale value of the biometric image, regions with grayscale values ​​of 0 or 255 are classified as background regions. This allows for rapid initial screening and classification of the image, excluding regions that are clearly not biometric, thereby improving the efficiency and accuracy of subsequent processing. Furthermore, this classification method does not require complex calculations and analysis, enabling region division in a short time and improving the overall processing speed and efficiency of the biometric recognition system.

[0068] In one possible implementation, the division of the biometric image into background and foreground regions includes: for regions in the biometric image where there are no grayscale values ​​of 0 or 255, converting the grayscale values ​​of the pixels in the biometric image into corresponding signal quantities; and dividing the pixels of the biometric image into the background and foreground regions respectively based on the signal quantities.

[0069] Specifically, after classifying areas with grayscale values ​​of 0 or 255 as background regions, the remaining areas of the biometric image can be segmented. First, the grayscale values ​​of the pixels in the biometric image can be mapped back to the original physical signal quantities. These signal quantities reflect the magnitude of physical quantities such as light intensity or capacitance changes received by the image sensor. In biometric images, areas with larger signal quantities typically indicate locations where biometric features are present and clearly visible, while areas with smaller signal quantities may belong to the background.

[0070] Since grayscale values ​​can be generated when a sensor converts physical signals into digital signals during the acquisition of biometric images, the grayscale values ​​can be reverse-mapped back to the original physical signals, and the foreground and background regions can be divided based on the amount of physical signals.

[0071] In this embodiment of the disclosure, by converting grayscale values ​​into signal quantities and dividing regions accordingly, the background and foreground regions can be identified more accurately, reducing misclassification and providing a more accurate image basis for subsequent operations such as biometric extraction and outlier detection, thereby improving the accuracy and reliability of the entire biometric recognition process.

[0072] In one possible implementation, the step of dividing the pixels of the biometric image into the background region and the foreground region according to the semaphore includes: sorting the semaphore according to a preset arrangement rule; selecting a first reference value from the sorted semaphore according to a first preset ratio; multiplying the first reference value by a second preset ratio to obtain a classification threshold; and classifying the pixels corresponding to semaphores smaller than the classification threshold into the background region, and classifying the remaining pixels into the foreground region.

[0073] Specifically, after converting the grayscale values ​​of pixels in a biometric image into corresponding signal quantities, the signal quantities can be sorted based on a preset sorting rule. This preset sorting rule could be to sort the signal quantities from largest to smallest, thus obtaining the distribution of the signal quantities.

[0074] Furthermore, a biometric signal can be selected from the sorted signals according to a first preset ratio, serving as a first reference value. The first preset ratio is the proportion of the total number of pixels, used to locate the signal that is most likely to belong to the foreground region from the sorted signals.

[0075] Since the signal intensity in the background region is generally weak, even approaching zero, while the signal intensity in the biometric feature-covered region is much higher, the signal intensity distribution may exhibit two distinct peaks: one corresponding to the low signal intensity in the background region and the other corresponding to the high signal intensity in the biometric feature-covered region—a bimodal distribution. Based on this, according to a first preset ratio, a signal intensity can be selected from the sorted signal intensity; this signal intensity is located at or near the peak corresponding to the biometric feature-covered region. Therefore, this signal intensity selected based on the first preset ratio can be used as a benchmark value to ensure that effective biometric signal intensity is obtained.

[0076] Furthermore, the first baseline value and the second preset ratio can be multiplied to obtain a classification threshold. This classification threshold is used to distinguish between foreground and background regions. Pixels with signal values ​​less than the classification threshold are classified as background regions, while the remaining pixels are classified as foreground regions. This method allows for reasonable region division of biometric images. The second preset ratio can be approximately 50%, for example, its range can be 40%-60%. The actual values ​​of the first and second preset ratios can be determined based on the specific scenario, and this disclosure does not impose specific limitations on them.

[0077] For example, Figure 5 shows a schematic histogram of signal magnitude distribution corresponding to a biometric image according to an embodiment of the present disclosure. Referring to Figure 5, Figure A in Figure 5 shows a schematic diagram of the acquired biometric image, and Figure B in Figure 5 shows a schematic histogram of signal magnitude distribution corresponding to the biometric image in Figure A. In Figure B in Figure 5, the horizontal axis represents the signal magnitude, and the vertical axis represents the number of pixels corresponding to the signal magnitude. Starting from the right side of the horizontal axis, the signal magnitudes are arranged in descending order. It can be seen that the maximum signal magnitude is 50. Assuming that the first preset ratio is 20%, then the first benchmark value is the signal value corresponding to 20% of the total number of pixels, starting from 50. Here, the signal value is 40, and this value is used as the first benchmark value (i.e., the black line at the horizontal axis value of 40). Assuming that the second preset ratio is 50%, then multiplying 40 by 50%, i.e., the classification threshold is 20 (i.e., the black line at the horizontal axis value of 20). Then, the pixels to the left of 20 are classified as the background area, and the remaining pixels are classified as the foreground area.

[0078] Figure 6 shows a histogram of signal magnitude distribution corresponding to another biometric image according to an embodiment of the present disclosure. Referring to Figure 6, Figure A in Figure 6 shows a schematic diagram of the acquired biometric image, and Figure B in Figure 6 shows a histogram of signal magnitude distribution corresponding to the biometric image in Figure A. The horizontal and vertical axes in Figure B in Figure 6 are the same as those in Figure B in Figure 5. Referring to Figure 6, it can be seen that when biometric features are acquired, if the acquisition area of ​​the capacitive module is mostly or completely covered (i.e., Figure A in Figure 6), then, according to the classification threshold selected based on the first preset ratio, the reference value, and the second preset ratio (20 in Figure B in Figure 6), all effective signal magnitudes can be assigned to the foreground region. Therefore, by setting the first preset ratio, the first reference value, and the second preset ratio, the effective division of the foreground and background regions can be achieved, thereby improving the accuracy of outlier identification.

[0079] In this embodiment, the distribution patterns of signal quantities are utilized to divide the foreground and background regions. Specifically, the signal quantities are sorted according to a preset arrangement rule to obtain their distribution characteristics. Then, a reference value is selected from the sorted signal quantities using a first preset ratio. This value represents the median value of effective biometric signals. The reference value is then multiplied by a second preset ratio to generate a classification threshold, which serves as the boundary between the background and foreground regions, ensuring that effective signals of moderate intensity are not missed as much as possible. Pixels corresponding to signal quantities smaller than the classification threshold are then assigned to the background region, and the remaining pixels are assigned to the foreground region. This achieves accurate division of the foreground and background regions, providing a reliable data foundation for subsequent anomaly detection. The first and second preset ratios can be dynamically adjusted based on actual needs. This dynamic adjustment mechanism can adapt to different biometric coverage levels (such as complete coverage or partial contact) and differences in sensor characteristics, avoiding the limitations of fixed thresholds in complex scenarios. For example, the second ratio can be reduced in noisy environments to enhance anti-interference capabilities, or the ratio can be increased to expand the foreground coverage when high sensitivity is required.

[0080] S12. Construct a statistical data structure that matches the pixels of the biometric image, and determine the value of the statistical data structure based on the results of the region division;

[0081] A statistical data structure is a storage unit used to record the frequency of each pixel in a biometric image being divided into background regions. The statistical data structure can be a data organization method designed for storing, processing, and analyzing statistical information. The statistical data structure can be a matrix, array, hash table, trie, or other data structures, and its specific form can be designed according to actual needs; this disclosure does not impose any limitations on this. The initial value of the statistical data structure can be filled with any numerical value; for example, it can be filled with 0, and this disclosure does not impose any limitations on this.

[0082] In one possible implementation, constructing a statistical data structure that matches the pixels of the biometric image includes: obtaining the resolution of the biometric image; and constructing a statistical data structure with the same size as the resolution.

[0083] Specifically, the resolution of the biometric image (i.e., the width and height of the image) can be obtained, and a statistical data structure that is completely consistent with the resolution can be constructed, so that the position of each pixel has a unique mapping in the data structure. For example, when the resolution of the biometric image is M×N, the size of the statistical data structure is also M×N, and each element corresponds to the statistical value of a pixel.

[0084] For example, assuming the image acquired by the capacitor module is a biometric image with a resolution of 3×3, and the statistical data structure is a matrix, then the matrix size is 3×3, and the initial values ​​of the matrix are filled with 0. Then the constructed matrix A can be expressed by formula (1):

[0085] (1)

[0086] In matrix A, the first row represents the first row of pixels in the biometric image, the second row represents the second row of pixels in the biometric image, and the third row represents the third row of pixels in the biometric image.

[0087] In this embodiment of the disclosure, by acquiring the resolution of the biometric image and then constructing a statistical data structure with the same size as the resolution, since the size of the statistical data structure is consistent with the resolution, it can accurately correspond to each pixel. This allows for full utilization of various types of data recorded in the statistical data structure, such as the original data of each pixel and the relevant feature parameters statistically derived from these original data, when performing pixel-based data processing operations such as feature extraction and image enhancement on the biometric image, thereby improving the accuracy of data processing.

[0088] For example, in the subsequent outlier identification process, the numerical values ​​of different pixels can be obtained more accurately, thereby improving the accuracy of outlier identification. At the same time, this targeted data structure also provides a more efficient data management mode for the entire biometric image processing system, reducing unnecessary data redundancy and complex calculations during processing, improving the overall operating efficiency of the system, and helping to achieve the goal of fast and accurate biometric image processing.

[0089] In one possible implementation, determining the value of the statistical data structure based on the result of region division includes: increasing the value of the statistical data structure corresponding to the background region by a first preset value, and decreasing the value of the statistical data structure corresponding to the foreground region by a second preset value, wherein the absolute value of the first preset value is less than the absolute value of the second preset value.

[0090] After constructing the statistical data structure, values ​​can be assigned to the statistical data structure based on the segmentation results of the foreground and background regions of the biometric image. Specifically, based on the initial values ​​of the statistical data structure, the values ​​of the statistical data structure corresponding to pixels segmented into the background region are increased by a first preset value, and the values ​​of the statistical data structure corresponding to pixels segmented into the foreground region are decreased by a second preset value, wherein the absolute value of the first preset value is less than the absolute value of the second preset value.

[0091] For example, Figure 7 shows a schematic diagram of the result of dividing the background and foreground regions of a biometric image according to an embodiment of the present disclosure. As shown in Figure 7, assuming that the white area is divided into the background region and the black area is divided into the foreground region, and assuming that the first preset weight is 1 and the second preset weight is 2, then matrix A can be expressed by formula (2):

[0092] (2)

[0093] In this embodiment, since the absolute value of the first preset value is less than the absolute value of the second preset value, if a pixel is classified as background, the statistical value increases by the first preset value (e.g., +1). This small increment allows for the careful accumulation of statistical values ​​for potential outliers, preventing a rapid increase in statistical values ​​due to a single misjudgment. If a pixel is classified as foreground, the statistical value decreases by the second preset value (e.g., -2). This larger decrease quickly offsets any previous possible accumulation of misjudgments, preventing edge areas (such as incomplete fingerprint coverage in a fingerprint image) from being misjudged as outliers due to occasional misclassification as background.

[0094] S13. When the biometric image is updated, the numerical value of the statistical data structure is updated, wherein the numerical value represents the frequency at which each pixel in the biometric image is divided into a background region.

[0095] When the capacitor module acquires a new image, i.e. when the biometric image is updated, outliers may appear due to capacitor plate failure, short circuit in sensor circuits, etc., i.e., hardware-related outliers. When the biometric image is updated, the position of the outliers does not change. Therefore, by using the changes in the numerical values ​​of the statistical data structure, we can accurately and intuitively reflect the frequency at which each pixel in the biometric image is classified as a background region. Pixels that have been classified as background regions for a long time after multiple image updates can be identified as outliers.

[0096] In one possible implementation, before or after updating the biometric image, if a second value in the statistical data structure is less than the initial value, the second value can be adjusted to the initial value. This avoids the need for multiple image updates before a new outlier can be identified.

[0097] For example, after assigning matrix A based on the segmentation results of the biometric image in Figure 7, before updating the biometric image, the values ​​in matrix A that are less than the initial value can be adjusted to the initial value, which is 0. Therefore, matrix A can be expressed by formula (3):

[0098] (3)

[0099] In one possible implementation, when the biometric image is updated, the numerical update of the statistical data structure includes: when the biometric image is updated, updating the numerical values ​​of the statistical data structure corresponding to the pixels in the background region and the foreground region according to the first preset value and the second preset value.

[0100] Specifically, after the capacitor module acquires a new image, i.e., during biometric image updates, the background and foreground regions are redefined after the biometric image update. Based on the updated biometric image region division results, the values ​​of the statistical data structure corresponding to the background region are increased by a first preset value, and the values ​​of the statistical data structure corresponding to the foreground region are decreased by a second preset value. This achieves the updating of the statistical data structure values ​​corresponding to the pixels in the background and foreground regions. By performing asymmetrical value updates to the statistical data structure, the misjudgment of an area as an outlier due to multiple instances of incomplete or absent signal coverage in the edge region during biometric image acquisition can be effectively avoided.

[0101] For example, Figure 8 shows a schematic diagram of the result of background and foreground region division of the updated biometric image according to an embodiment of the present disclosure. Referring to Figure 8, continuing to assume that the white area is divided into the background region and the black area is divided into the foreground region, after the biometric image is updated to Figure 8, matrix A can be expressed by formula (4):

[0102] (4)

[0103] Among them, the pixels that are divided into the foreground region are subjected to a -2 operation, the pixels that are divided into the background region are subjected to a +1 operation, and the value of the statistical data structure corresponding to the pixels that are divided into the foreground region is less than the initial value 0, so it can be restored to the initial value, thus obtaining matrix A in formula (4).

[0104] In this embodiment of the disclosure, when updating the biometric image, the statistical values ​​can dynamically accumulate the classification tendency of the pixels through a differentiated adjustment mechanism of the first preset value and the second preset value. Pixels that are frequently classified as background will have their statistical values ​​gradually increase and eventually exceed the preset threshold to be judged as outliers. Meanwhile, edge areas that are occasionally misclassified as background will quickly return to zero due to the large reduction (e.g., -2) when dividing the foreground, thus avoiding misclassification.

[0105] S14. If the statistical data structure contains a first value that reaches a first preset threshold, the biometric image pixel corresponding to the first value is identified as an outlier.

[0106] Specifically, the larger the value of the statistical data structure, the higher the frequency with which the pixel is continuously classified into the background region, meaning it is more likely to be an outlier. When the statistical data structure has a first value that reaches a first preset threshold, the biometric image pixel corresponding to the first value can be identified as an outlier. The actual value of the first preset threshold can be determined according to actual needs, and this disclosure does not impose specific limitations on it.

[0107] For example, suppose that the capacitor module has collected 10 biometric images, that is, collected 10 different biometric images. The division of the biometric image regions collected in the first and second times can be seen in Figures 7 and 8. After 10 iterations of biometric images, suppose that matrix A has iterated into the matrix shown in formula (5). Suppose that the first preset threshold is 10. Then it can be seen that the value that reaches the threshold is A

[21] , that is, the first pixel in the second row. Then the first pixel in the second row of the biometric image can be identified as the outlier.

[0108] (5)

[0109] It should be noted that the 3×3 resolution biometric image used in this embodiment is only a simplified example. In actual application scenarios, the resolution and size of biometric images commonly range from hundreds of thousands of pixels to millions of pixels, and the method in this embodiment is also applicable.

[0110] In this embodiment, a biometric image is acquired and divided into background and foreground regions. A statistical data structure matching the pixels of the biometric image is constructed, and the value of the statistical data structure is determined based on the region division result. When the biometric image is updated, the value of the statistical data structure is updated, whereby the value represents the frequency with which each pixel in the biometric image is classified as a background region. When a first value in the statistical data structure reaches a first preset threshold, the biometric image pixel corresponding to the first value is identified as an outlier. Thus, outliers can be effectively and accurately identified in the biometric image.

[0111] Application scenario examples

[0112] This disclosure includes an application example, which includes the following:

[0113] 1. The input image is first divided into regions. At this time, both the real background region and the outlier will be marked as the background region. Then, the outlier is distinguished by dynamic statistics.

[0114] 2. There are two methods for dividing the foreground and background areas in step 1:

[0115] For images with saturation, all areas with a grayscale value of 0 or 255 are identified as background areas.

[0116] For images without saturation, the grayscale images are first restored to their signal magnitudes and sorted from largest to smallest. The signal magnitude at 20% of the signal value sequence (refer to the first vertical line from right to left on the horizontal axis in Figure 5 and Figure 6) is taken as the fingerprint signal magnitude (V0). Furthermore, 50% of this (refer to the second vertical line from right to left on the horizontal axis in Figure 5 and Figure 6) is taken as the threshold for distinguishing between blank areas and foreground areas (thre = 0.5*V0). Areas smaller than this threshold are determined to be background areas.

[0117] 3. Set a background area statistics array with the same size as the image. This statistics array will be updated every time the fingerprint image is updated.

[0118] 4. The array starts counting from 0. When a point in the image is classified as a background region, the count is incremented by 1; when a point is classified as a foreground region, the count is decremented by 2 (if the number obtained after each array update is negative, it is reset to 0). After multiple updates, the pixels corresponding to positions in the entire array with a statistical value greater than 10 (represented as background regions for a long time) are identified as outliers.

[0119] The purpose of varying the size of the fingerprints is to prevent edge areas from being frequently not covered during continuous fingerprint recording, which could lead to misidentification as outliers.

[0120] Figure 9 illustrates a fingerprint image region division and local magnification according to an embodiment of the present disclosure. Referring to Figure 9, in step 1, the input image is shown in Figure 9A. In step 2, the image obtained after dividing the foreground / background is shown in Figure 9B. In step 4, the area divided into a background region is the white area in Figure 9B, and the area divided into a foreground region is the black area in Figure 9B. The divided regions are then locally magnified, as shown in Figure 9C. After local magnification, the corresponding array increases or decreases are shown in Figure 9C.

[0121] In the embodiments disclosed herein, a continuous statistical method is used to continuously collect data on the pixels of the capacitor module, thereby dynamically identifying defective pixels in the capacitor module. This process does not rely on specialized detection tools. In practical applications, image processing techniques, such as inpainting, can be used to repair biometric images containing defective pixels, thereby effectively reducing the adverse effects of defective pixels on the accuracy of biometric recognition, improving module yield, and extending module lifespan.

[0122] Figure 10 shows a block diagram of an image outlier detection apparatus according to an embodiment of the present disclosure. As shown in Figure 10, the apparatus includes:

[0123] Image segmentation module 21 is used to acquire biometric images and segment the biometric images into background and foreground regions;

[0124] The statistical data structure construction module 22 is used to construct a statistical data structure that matches the pixels of the biometric image, and to determine the value of the statistical data structure based on the result of the region division.

[0125] The numerical update module 23 is used to update the numerical values ​​of the statistical data structure when the biometric image is updated, wherein the numerical values ​​represent the frequency at which each pixel in the biometric image is divided into a background region.

[0126] The outlier determination module 24 is used to determine the biometric image pixel corresponding to the first value as an outlier when the statistical data structure has a first value that reaches a first preset threshold.

[0127] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0128] For example, the electronic devices in this embodiment include, but are not limited to, desktop computers, televisions, mobile devices with large screens such as mobile phones and tablets, and other common electronic devices that require multiple chips to be cascaded together to achieve driving.

[0129] For example, electronic devices can also be user equipment (UE), mobile devices, user terminals, terminals, handheld devices, computing devices, or in-vehicle devices, etc. Examples of terminals include: displays, smartphones or portable devices, mobile phones, tablets, laptops, PDAs, mobile internet devices (MIDs), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, and wireless terminals in vehicle-to-everything (V2X) networks, etc. For example, a server can be a local server or a cloud server.

[0130] Figure 11 shows a block diagram of an electronic device 1900 according to an embodiment of the present disclosure. For example, the electronic device 1900 may be provided as a server or a terminal device. Referring to Figure 11, the electronic device 1900 includes a processing component 1922, which further includes one or more processors, and memory resources represented by memory 1932 for storing instructions executable by the processing component 1922, such as application programs. The application programs stored in memory 1932 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 1922 is configured to execute instructions to perform the methods described above.

[0131] Electronic device 1900 may also include a power supply component 1926 configured to perform power management of electronic device 1900, a wired or wireless network interface 1950 configured to connect electronic device 1900 to a network, and an input / output interface 1958. Electronic device 1900 can operate on an operating system stored in memory 1932, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, or similar.

[0132] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions that can be executed by a processing component 1922 of an electronic device 1900 to perform the above-described method.

[0133] According to another aspect of this disclosure, a computer program product is provided, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device performs the above-described method.

[0134] According to another aspect of this disclosure, a chip is provided that includes the image outlier detection device described above.

[0135] For specific embodiments of the above-mentioned devices, electronic devices, and chips, please refer to the Method section.

[0136] The above description is merely an exemplary embodiment of the present invention and is not intended to limit the scope of protection of the present invention, which is determined by the appended claims.

[0137] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0138] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0139] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0140] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, and are not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for detecting outliers in an image, characterized in that, include: A biometric image is acquired, and the biometric image is divided into background and foreground regions. A statistical data structure matching the pixels of the biometric image is constructed, and the value of the statistical data structure is determined based on the region division result. When the biometric image is updated, the value of the statistical data structure is updated, and the value represents the frequency at which each pixel in the biometric image is divided into the background region. When a first value in the statistical data structure reaches a first preset threshold, the biometric image pixel corresponding to the first value is identified as an outlier. The step of determining the value of the statistical data structure based on the region division result includes: increasing the value of the statistical data structure corresponding to the background region by a first preset value, and decreasing the value of the statistical data structure corresponding to the foreground region by a second preset value, wherein the absolute value of the first preset value is less than the absolute value of the second preset value; the step of constructing a statistical data structure matching the pixels of the biometric image includes: obtaining the resolution of the biometric image; constructing a statistical data structure with the same size as the resolution; the step of updating the value of the statistical data structure when the biometric image is updated includes: when the biometric image is updated, re-dividing the background region and foreground region of the updated biometric image, and updating the value of the statistical data structure corresponding to the pixels of the background region and foreground region according to the first preset value and the second preset value.

2. The method according to claim 1, characterized in that, The step of dividing the biometric image into background and foreground regions includes: obtaining the grayscale value of the biometric image; and for regions in the biometric image where the grayscale value is 0 or 255, dividing the regions with grayscale values ​​of 0 or 255 into the background region.

3. The method according to claim 2, characterized in that, The step of dividing the biometric image into background and foreground regions includes: for regions in the biometric image where there are no grayscale values ​​of 0 or 255, converting the grayscale values ​​of the pixels in the biometric image into corresponding signal quantities; and dividing the pixels of the biometric image into the background and foreground regions respectively based on the signal quantities.

4. The method according to claim 3, characterized in that, The step of dividing the pixels of the biometric image into the background region and the foreground region according to the signal quantity includes: sorting the signal quantity according to a preset arrangement rule; selecting a first reference value from the sorted signal quantity according to a first preset ratio; multiplying the first reference value by a second preset ratio to obtain a classification threshold; and dividing the pixels corresponding to the signal quantity smaller than the classification threshold into the background region, and dividing the remaining pixels into the foreground region.

5. An image outlier detection device, characterized in that, The device includes: an image segmentation module for acquiring a biometric image and segmenting the biometric image into background and foreground regions; a statistical data structure construction module for constructing a statistical data structure matching the pixels of the biometric image and determining the value of the statistical data structure based on the region segmentation result; a value update module for updating the value of the statistical data structure when the biometric image is updated, wherein the value represents the frequency at which each pixel in the biometric image is segmented into a background region; and an outlier determination module for determining the biometric image pixel corresponding to the first value as an outlier when the statistical data structure has a first value that reaches a first preset threshold. The statistical data structure construction module is used to: increase the value of the statistical data structure corresponding to the background region by a first preset value, and decrease the value of the statistical data structure corresponding to the foreground region by a second preset value, wherein the absolute value of the first preset value is less than the absolute value of the second preset value; the statistical data structure construction module is used to: obtain the resolution of the biometric image; and construct a statistical data structure with the same size as the resolution; the value update module is used to: when the biometric image is updated, re-divide the background region and foreground region of the updated biometric image, and update the values ​​of the statistical data structure corresponding to the pixels in the background region and foreground region according to the first preset value and the second preset value.

6. A display device, characterized in that, It includes multiple display units and at least one image outlier detection device according to claim 5.

7. The display device according to claim 6, characterized in that, The display unit includes a display panel, which includes at least one of the following: liquid crystal display panel, micro light-emitting diode display panel, light-emitting diode display panel, mini light-emitting diode display panel, quantum dot light-emitting diode display panel, organic light-emitting diode display panel, cathode ray tube display panel, digital light processing display panel, field emission display panel, plasma display panel, electrophoretic display panel, electrowetting display panel, and small-pitch display panel.

8. An electronic device, characterized in that, Includes the display device according to claim 6 or 7.

9. An electronic device, characterized in that, include: processor; A memory for storing processor-executable instructions; wherein the processor is configured to implement the method of any one of claims 1 to 4 when executing the instructions stored in the memory.

10. A non-volatile computer-readable storage medium storing computer program instructions thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 4.

11. A chip, characterized in that, The chip includes the image outlier detection device as described in claim 5.

Images