Image processing method and apparatus, chip, storage medium, and electronic device

Abstract

The present application provides an image processing method and apparatus, a chip, a storage medium, and an electronic device. The method comprises: selecting, from a blank image library, a matching blank image that matches a blank image to be tested, to obtain an image to be determined; and when the image to be determined does not have background texture, and / or when the image to be tested has background texture but meets a preset condition, adding the blank image to the blank image library. Embodiments of the present application can improve the accuracy of fingerprint identification by means of the blank image library.

Description

[0001] Image Processing Method, Device, Chip, Storage Medium, and Electronic Device This application claims priority to a Chinese application filed with the China Patent Office on February 8, 2024, with application number 2024101782870, entitled "Image Processing Method, Device, Chip, Storage Medium, and Electronic Device," the entire contents of which are incorporated herein by reference. Technical Field This application relates to the field of finger-level recognition technology, and more particularly to an image processing method, device, chip, storage medium, and electronic device. Background: Currently, ultrasonic finger-level sensors (a type of chip) generate fixed signals (i.e., fixed noise) when capturing images. This fixed signal is superimposed on the finger-level signal, making it difficult to recover the finger-level path through preprocessing. Fixed signals include stripe-level noise generated at the edge of the image due to acoustic wave diffraction, fixed bad pixels, and the like. Related technologies typically acquire fixed signals by using a base image (an image sampled by the sensor, i.e., a base image, which may contain fixed-level noise from the edges of the ultrasonic array and fixed-level noise from electronic devices) to remove the fixed signals and recover the finger-level path. However, the acquisition of the base image (i.e., the base image) in related technologies is susceptible to factors such as the acquisition time and the screen conditions of the electronic device. This can cause the base image to contain a low-level signal, which can be superimposed on the finger-level path during base correction (i.e., removing fixed noise using the base image), resulting in anomalies in the recovered finger-level image. Therefore, there is currently no effective solution for image processing to reduce the impact of the low-level signal on the finger-level image during the fixed-noise removal process, thereby improving the accuracy of finger-level recognition.SUMMARY OF THE INVENTION In view of this, embodiments of the present application provide an image processing method, apparatus, chip, storage medium, and electronic device to address the problem that the air-sampled images acquired in related technologies may contain strong bottom-level signals, thereby affecting the correction effect of the finger-level image. Specifically, embodiments of the present application can perform image processing on the air-sampled image to be detected using a matching air-sampled image in an air-sampled image library that matches the air-sampled image to be detected, thereby obtaining a highly accurate image to be determined. The bottom level of the air-sampled image to be detected can be detected using the image to be determined, thereby obtaining an air-sampled image with no bottom level or a small bottom level signal (i.e., an air-sampled image to be detected when the image to be determined has no bottom level and / or has a bottom level but meets preset conditions). The air-sampled image library is then updated, and the air-sampled images in the air-sampled image library can be air-sampled images with no bottom level or a small bottom level signal. Furthermore, the air-sampled image library supports correction of finger-level images. The air-sampled images in the air-sampled image library can then be used to improve the correction effect of the finger-level image, thereby reducing the impact of the bottom level signal on the finger-level image during the fixed noise removal process. Thereby improving the accuracy of finger-level recognition. According to one aspect of an embodiment of the present application, an image processing method is provided, the method comprising: obtaining an empty-sampled image to be detected; selecting a matching empty-sampled image that matches the empty-sampled image to be detected from an empty-sampled image library, and using the matching empty-sampled image to remove fixed noise from the empty-sampled image to be detected to obtain an image to be determined; when the image to be determined has no bottom level, adding the empty-sampled image to be detected to the empty-sampled image library; and / or, when the image to be determined has a bottom level but meets a preset condition, adding the empty-sampled image to be detected to the empty-sampled image library, the empty-sampled image library supports correction of finger-level images.According to another aspect of an embodiment of the present application, an image processing device is provided. The device includes: an acquisition unit for acquiring a blank-sampled image to be detected; a processing unit for selecting a matching blank-sampled image that matches the blank-sampled image to be detected from a blank-sampled image library, and using the matching blank-sampled image to remove fixed noise from the blank-sampled image to obtain an image to be determined; the processing unit is further configured to add the blank-sampled image to be detected to the blank-sampled image library when the image to be determined does not have a bottom level; and / or to add the blank-sampled image to the blank-sampled image library when the image to be determined does have a bottom level but meets a preset condition. The blank-sampled image library supports correction of finger-level images. According to another aspect of an embodiment of the present application, a chip is provided. The chip is located in an electronic device and is configured to perform the aforementioned method. According to another aspect of an embodiment of the present application, an electronic device is provided. The electronic device includes one or more processors and one or more memories storing a program, wherein the program includes instructions that, when executed by the one or more processors, cause the one or more processors to perform the aforementioned method. According to another aspect of an embodiment of the present application, a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the aforementioned method is provided. In this embodiment of the present application, after acquiring a blank-sampled image to be detected, a matching blank-sampled image that matches the blank-sampled image to be detected is selected from a blank-sampled image library. The matching blank-sampled image can then be used to remove fixed noise from the blank-sampled image to be detected, yielding an image to be determined. This removes fixed signals (i.e., fixed noise) from the blank-sampled image to be detected, eliminating the fixed signals. Furthermore, the fixed signals in the matching blank-sampled image are more similar to those in the blank-sampled image to be detected, thereby improving the accuracy of the image to be determined. Accordingly, when the image to be determined has no bottom level, the blank-sampled image to be detected is added to the blank-sampled image library; and / or, when the image to be determined has a bottom level but meets preset conditions, the blank-sampled image to be detected is added to the blank-sampled image library. The blank-sampled image library supports the correction of finger-level images.It can be seen that the embodiment of the present application can perform image processing on the empty-sampled image to be detected through the matching empty-sampled image in the empty-sampled image library that matches the empty-sampled image to be detected, so as to obtain a more accurate image to be determined, so as to realize the bottom-level detection of the empty-sampled image to be detected through the image to be determined, thereby obtaining an empty-sampled image with no bottom level or a small bottom-level signal (that is, the empty-sampled image to be detected when the image to be determined has no bottom level and / or the image to be determined has a bottom level but meets the preset conditions), and then update the empty-sampled image library. Then, the empty-sampled images in the empty-sampled image library can be empty-sampled images with no bottom level or a small bottom-level signal, and the empty-sampled image library supports correction of finger-level images. Then, the empty-sampled images in the empty-sampled image library can be used to improve the finger-level image correction effect, so as to reduce the influence of the bottom-level signal on the finger-level image during the fixed noise removal process, thereby improving the accuracy of finger-level recognition. BRIEF DESCRIPTION OF THE DRAWINGS In the following description of exemplary embodiments in conjunction with the accompanying drawings, further details, features, and advantages of the present application are disclosed. In the accompanying drawings: FIG1 is a schematic flow chart of an image processing method according to an exemplary embodiment of the present application; FIG2 is a schematic flow chart of another image processing method according to an exemplary embodiment of the present application; FIG3 is a schematic flow chart of yet another image processing method according to an exemplary embodiment of the present application; FIG4 is a schematic block diagram of an image processing apparatus according to an exemplary embodiment of the present application; and FIG5 is a block diagram of an exemplary electronic device capable of implementing the embodiments of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application may be implemented in various forms and should not be construed as limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are for illustrative purposes only and are not intended to limit the scope of protection of the present application. It should be understood that the various steps described in the method embodiments of the present application may be performed in a different order and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit illustrated steps. The scope of the present application is not limited in this respect. The term "include" and its variations used herein are open inclusions, i.e., "including but not limited to".The term "based on" means "based at least in part on." The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; and the term "some embodiments" means "at least some embodiments." Definitions of other terms are provided below. It should be noted that the terms "first" and "second," etc., used herein, are intended solely to distinguish different devices, modules, or units and are not intended to limit the order or interdependencies of the functions performed by these devices, modules, or units. It should be noted that the modifiers "one" and "a plurality" used herein are illustrative and non-restrictive. Those skilled in the art will understand that, unless the context clearly indicates otherwise, they should be understood to mean "one or more." The names of the messages or information exchanged between the multiple devices in the embodiments of this application are for illustrative purposes only and are not intended to limit the scope of these messages or information. It should be noted that the image processing method provided in the embodiments of the present application may be executed by an electronic device; alternatively, the electronic device may include a chip, and the image processing method may also be executed by the chip in the electronic device, which is not limited in the embodiments of the present application. Optionally, the chip in the electronic device may be an ultrasonic finger-level sensor, which can be used to capture images, such as finger-level images or air-sampled images. The electronic device may be a terminal (i.e., a client) or a server. Accordingly, the terminals mentioned herein may include, but are not limited to, smartphones, wristbands, tablets, laptops, desktop computers, smartwatches, intelligent voice interaction devices, and the like. The server mentioned herein may be an independent physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.Based on the above description, embodiments of the present application provide an image processing method. This image processing method can be performed by the aforementioned electronic device or by a chip within the electronic device. For ease of illustration, the following description uses the electronic device performing the image processing method as an example. As shown in FIG1 , the image processing method may include the following steps S101-S103:

[0002] S101: Acquire a void-collected image to be detected. A void-collected image may also be referred to as an ultrasonic base image or base image. Optionally, there may be one or more void-collected images to be detected, which is not limited in this embodiment of the present application. When there are multiple void-collected images to be detected, the electronic device may perform image processing on each of the multiple void-collected images to achieve bottom-level detection of each void-collected image, thereby updating a void-collected image library. For ease of illustration, the following description uses a single void-collected image to be detected as an example. It should be understood that the void-collected image library may include at least one void-collected image. In this embodiment of the present application, the methods for acquiring the void-collected image to be detected may include, but are not limited to, the following: First Acquisition Method: The electronic device may include a chip (i.e., an ultrasonic finger-level sensor). In this case, the electronic device may acquire the void-collected image to be detected using the ultrasonic finger-level sensor to acquire the void-collected image to be detected. Optionally, the collected empty-sampled image to be detected can also be referred to as a real-time base. Optionally, the empty-sampled image to be detected can be collected within a preset collection time period starting from the time the user lifts their finger from the collection area. This allows the collection time of the empty-sampled image to be detected to be similar to the collection time of the finger-level image. This allows for the partial image indication data (such as the acquisition temperature and acquisition frequency) in the image indication dataset of the empty-sampled image to be detected to be similar to the partial image indication data in the image indication dataset of the finger-level image, thereby further aligning the fixed signals in the empty-sampled image to be detected with the fixed signals in the empty-sampled image to be detected. Alternatively, the empty-sampled image to be detected can be collected when the user's finger is not touching the collection area within a specified time range, and so on. This embodiment of the present application is not limited in this regard. Clearly, the empty-sampled image to be detected can be collected when the user's finger is not touching the collection area. Optionally, both the preset collection time period and the specified time range can be set based on experience or actual needs, and this embodiment of the present application is not limited in this regard. The second acquisition method: The electronic device may store multiple undetected empty sampling images in its own storage space. In this case, the electronic device may select one undetected empty sampling image from the multiple undetected empty sampling images and use the selected undetected empty sampling image as the empty sampling image to be detected.Third acquisition method: The electronic device can obtain a download link for an uncollected image. In this case, the uncollected image downloaded from the download link can be used as the uncollected image to be detected, and so on. S102: A matching uncollected image that matches the uncollected image to be detected is selected from the uncollected image library. The matching uncollected image is used to remove fixed noise from the uncollected image to be detected, thereby obtaining an image to be determined. Optionally, the uncollected image library can also be referred to as a base library. The electronic device can search the base library to obtain an optimal cached base (i.e., a matching uncollected image). Optionally, the electronic device can include an uncollected image correction module (i.e., a base correction module). In this case, the electronic device can obtain a matching uncollected image through the base correction module and use the matching uncollected image to remove fixed noise from the uncollected image to be detected, thereby obtaining an image to be determined. This corrects the uncollected image to ensure that the image to be determined does not contain fixed noise. Specifically, the electronic device may determine an image indication dataset of a to-be-detected empty image, select M target empty images from the empty image library, and determine an image indication dataset for each of the M target empty images. The image indication dataset for each image may include at least one of the following: an image grayscale matrix of the corresponding image (which may be used to describe level information), an acquisition frequency of the corresponding image (i.e., the frequency of the ultrasonic wave when the corresponding image was acquired), an acquisition temperature of the corresponding image (i.e., the temperature of the electronic device when the corresponding image was acquired), an acquisition time of the corresponding image (i.e., the time when the corresponding image was acquired), and a flight time of the corresponding image (i.e., the time interval between the signal transmission time and the signal reception time of the ultrasonic level sensor when acquiring the corresponding image). Specifically, the acquisition frequency of the corresponding image is the signal transmission frequency of the ultrasonic signal that generates the image (i.e., the corresponding image), and the flight time of the corresponding image is the signal flight time of the ultrasonic signal that generates the image (i.e., the corresponding image). M is a positive integer; it should be understood that M is less than or equal to the number of empty images in the empty image library. Optionally, the air-sampled image library may include an image indication dataset of each air-sampled image in at least one air-sampled image, that is, may include an image indication dataset of each target air-sampled image.It should be understood that the closer the acquisition frequency of the matching empty-sampled image is to the acquisition frequency of the empty-sampled image to be detected, the closer the fixed noise signal (i.e., fixed noise or fixed signal) of the matching empty-sampled image can be to the fixed noise signal of the empty-sampled image to be detected, and the better the base correction (i.e., empty-sampled image correction) effect can be. Correspondingly, the closer the empty-sampled image to be detected is to the matching empty-sampled image, the closer the fixed noise signal of the matching empty-sampled image can be to the fixed noise signal of the empty-sampled image to be detected, and the better the base correction effect can be, and so on. In one embodiment, when selecting M target empty-sampled images from the empty-sampled image library, the electronic device can add each empty-sampled image in the empty-sampled image library to the M target empty-sampled images, so that the M target empty-sampled images include each empty-sampled image in the empty-sampled image library. In another embodiment, for any uncollected image in the uncollected image library, the electronic device may determine at least one image indication data from the image indication data set of any uncollected image; if the difference between each image indication data in the at least one image indication data and the corresponding image indication data of the uncollected image to be detected is less than a corresponding preset difference threshold, the uncollected image is added to M target uncollected images to obtain M target uncollected images. It should be understood that if the difference between the at least one image indication data and the corresponding image indication data of the uncollected image to be detected is greater than or equal to a corresponding preset difference threshold, the uncollected image is not added to the M target uncollected images. Optionally, the at least one image indication data may include, but is not limited to, the acquisition frequency and flight time of any uncollected image. Optionally, each image indication data in the at least one image indication data may correspond to a preset difference threshold, and the preset difference threshold corresponding to each image indication data in the at least one image indication data may be set based on experience or actual needs, which is not limited in this embodiment of the present application.It can be seen that when the difference between at least one image indication data and the corresponding image indication data of the uncollected image to be detected is too large, it is possible to avoid adding any of the above uncollected images to the M target uncollected images, thereby avoiding the subsequent selection of uncollected images with excessive differences as matching uncollected images. In addition, uncollected images with excessive differences from the uncollected images to be detected in the uncollected image library can be screened out in advance, thereby improving the accuracy of matching uncollected images and the efficiency of selecting matching uncollected images. For example, assuming the ultrasonic finger-level sensor has an acquisition frequency between 8MHz (megahertz) and 12MHz, at least one image indication data set includes the acquisition frequency of a corresponding empty-sampled image, the acquisition frequency of the empty-sampled image to be detected is 10MHz, and the preset difference threshold corresponding to the image indication data "acquisition frequency" is 1MHz. In this case, if the difference between the acquisition frequency of any empty-sampled image and the acquisition frequency of the empty-sampled image to be detected is less than 1MHz, then any empty-sampled image is added to the M target empty-sampled images. Specifically, empty-sampled images with acquisition frequencies between 9MHz and 11MHz are added to the M target empty-sampled images. Based on this, the electronic device can calculate the similarity between the image indication dataset of the empty-sampled image to be detected and the image indication datasets of each target empty-sampled image. The similarity between the image indication dataset of the empty-sampled image to be detected and the image indication dataset of a target empty-sampled image is used to indicate the similarity between the empty-sampled image to be detected and the corresponding target empty-sampled image.In a specific implementation, for any target uncollected image among the M target uncollected images and for any image indication data in the image indication data set of any target uncollected image, the electronic device may determine a similarity calculation method corresponding to any image indication data of any target uncollected image; and calculate, according to the similarity calculation method, the similarity under any image indication data of any target uncollected image based on any image indication data of any target uncollected image and the image indication data of the uncollected image to be detected, so as to obtain the similarity under each image indication data of any target uncollected image; the similarity under any image indication data of any target uncollected image may refer to: the similarity between any image indication data of any target uncollected image and the image indication data of the uncollected image to be detected; the type of the image indication data of the uncollected image to be detected (the type of one image indication data may also be referred to as the type of one image indication data) is the same as the type of any image indication data of any target uncollected image, that is, the image indication data of the uncollected image to be detected may refer to: image indication data of the same type as any image indication data in the image indication data set of the uncollected image to be detected; based on this, The similarity under each image indication data of any target empty-collected image may be weightedly summed to obtain the similarity between the image indication dataset of the empty-collected image to be detected and the image indication dataset of any target empty-collected image.Optionally, the weight value of the similarity under each image indication data may be set based on experience or actual needs, and this is not limited in the embodiments of the present application. Exemplarily, the weight value of the similarity under each image indication data of any target uncollected image may be, in descending order, the following: the similarity under the image grayscale matrix of any target uncollected image (i.e., the level similarity), the similarity under the acquisition frequency of any target uncollected image, the similarity under the acquisition temperature of any target uncollected image, the similarity under the flight time of any target uncollected image, and the similarity under the acquisition time of any target uncollected image. That is, when calculating the similarity between any target uncollected image and the uncollected image to be detected, the priority order of the used similarities may be, in order, the similarity under the image grayscale matrix of any target uncollected image, the similarity under the acquisition frequency of any target uncollected image, the similarity under the acquisition temperature of any target uncollected image, the similarity under the flight time of any target uncollected image, and the similarity under the acquisition time of any target uncollected image. It should be understood that the image indication data in the image indication data set of the empty-sampled image to be detected corresponds to the image indication data in the image indication data set of any target empty-sampled image, such as the acquisition frequency of the empty-sampled image to be detected corresponds to the acquisition frequency of any target empty-sampled image, the image grayscale matrix of the empty-sampled image to be detected corresponds to the image grayscale matrix of any target empty-sampled image, and so on; that is, one image indication data belongs to one type, and image indication data of the same type correspond to each other, and the image indication data set of one image may include at least one of the following types: image grayscale matrix type, acquisition frequency type, acquisition temperature type, acquisition time type, and flight time type, etc., such as the acquisition frequency of one image belongs to the acquisition frequency type (that is, the acquisition frequency of one image may be of the acquisition frequency type), the image grayscale matrix of one image belongs to the image grayscale matrix type (that is, the image grayscale matrix of one image may be of the image grayscale matrix type), and so on. Exemplarily, the image grayscale matrix type of the to-be-detected empty sampled image is the same as the image grayscale matrix type of any target empty sampled image, the acquisition frequency type of the to-be-detected empty sampled image is the same as the acquisition frequency type of any target empty sampled image, and so on.Optionally, if any image indication data (i.e., any image indication data of any target uncollected image) is an image grayscale matrix of any target uncollected image, then the similarity calculation method corresponding to any image indication data may be a structural similarity index measurement (SSIM) calculation method, or a cosine similarity measurement method, etc., which is not limited in this embodiment of the present application. For example, when any image indication data is an image grayscale matrix of any target uncollected image, and the similarity calculation method corresponding to any image indication data is a cosine similarity measurement method, the image indication data of the uncollected image to be detected may be the image grayscale matrix of the uncollected image to be detected. The electronic device may use Formula 1.1 based on any image indication data and the image indication data of the uncollected image to be detected. The i-th feature (i.e., the i-th pixel value) in the image grayscale matrix of the target empty-sampled image, B can represent the image grayscale matrix of the empty-sampled image to be detected, Bi can represent the i-th feature in the image grayscale matrix of the empty-sampled image to be detected, i is a positive integer less than or equal to n, and n can be the number of features included in an image (i.e., the number of features included in the image grayscale matrix of an image), i.e., the number of pixels in an image. Optionally, if the numerical values ​​of different feature dimensions differ significantly, a custom weight method can be used to perform feature similarity measurement calculations. That is, the electronic device can use Formula 1.2 to calculate any This can be set based on experience or actual needs, and is not limited in this embodiment of the present application. Accordingly, if any image indication data is a single feature such as the acquisition frequency or flight time of any target air-collected image, the similarity calculation method corresponding to any image indication data can be a reciprocal distance calculation method, for example. The reciprocal distance calculation method involves performing a reciprocal operation on the difference between two image indication data. For this purpose, using the example of any image indication data being the acquisition frequency of any target air-collected image, the electronic device can calculate the frequency difference between the acquisition frequency of any target air-collected image and the acquisition frequency of the air-collected image to be detected, and use the reciprocal of this frequency difference as the similarity for any image indication data. In this case, the smaller the frequency difference, the greater the similarity for any image indication data. In other words, the closer the acquisition frequencies, the greater the similarity. Furthermore, based on the similarity between the undetected image and each target undetected image, the electronic device may select the target undetected image with the greatest similarity to the undetected image among the M target undetected images as the matching undetected image. In other words, the matching undetected image may be selected from the M target undetected images to match the undetected image. The similarity between the undetected image and the matching undetected image is greater than or equal to the similarity between the undetected image and the target undetected images other than the matching undetected image among the M target undetected images. In a specific implementation, the electronic device may directly select the target undetected image with the greatest similarity to the undetected image from the M target undetected images to obtain the matching undetected image. In this case, when there is only one target uncollected image with the greatest similarity to the uncollected image to be detected, a matching uncollected image can be directly obtained; when there are multiple target uncollected images with the greatest similarity to the uncollected image to be detected, one target uncollected image can be randomly selected from the multiple target uncollected images with the greatest similarity to obtain a matching uncollected image; or, among the multiple target uncollected images with the greatest similarity, the target uncollected image with the greatest similarity at the acquisition frequency can be used as the matching uncollected image, and so on; this embodiment of the present application is not limited to this.In another specific implementation, the electronic device may sort the target air-collected images based on the similarity between the air-collected image to be detected and the target air-collected images (e.g., sorting in descending order of similarity, or sorting in ascending order of similarity, etc.), obtain a sorting result, and select a matching air-collected image from the sorting result. For example, using sorting in descending order of similarity as an example, the first target air-collected image in the sorting result may be selected as the matching air-collected image. Optionally, if multiple target air-collected images have the same similarity to the air-collected image to be detected, the multiple target air-collected images with the same similarity may be randomly sorted. Alternatively, sorting indication information for each target air-collected image may be determined based on the similarity of each target air-collected image in terms of acquisition frequency and / or flight time, and the multiple target air-collected images with the same similarity may be sorted based on the sorting indication information for each target air-collected image, and so on. Optionally, for any of the multiple target air-sampled images with the same similarity, the similarity of any of the same target air-sampled images under the acquisition frequency (such as the reciprocal of the frequency difference between the acquisition frequency of any of the same target air-sampled images and the acquisition frequency of the air-sampled image to be detected) can be used as the ranking indication information of any of the same target air-sampled images; or, the similarity of any of the same target air-sampled images under the flight time (such as the reciprocal of the time difference between the flight time of any of the same target air-sampled images and the flight time of the air-sampled image to be detected) can be used as the ranking indication information of any of the same target air-sampled images; or, a weighted sum (such as a sum operation or an average operation, etc.) can be performed on the similarity of any of the same target air-sampled images under the acquisition frequency and the similarity of any of the same target air-sampled images under the flight time to obtain the ranking indication information of any of the same target air-sampled images, and so on.For example, when sorting is performed in descending order of similarity and there are multiple target air-sampled images with the same similarity, the multiple target air-sampled images with the same similarity can be sorted in descending order of sorting indication information; when sorting is performed in descending order of similarity and there are multiple target air-sampled images with the same similarity, the multiple target air-sampled images with the same similarity can be sorted in descending order of sorting indication information, and so on.

[0003] S103: When the image to be determined has no bottom level, the uncollected image to be detected is added to the uncollected image library; and / or, when the image to be determined has a bottom level but meets a preset condition, the uncollected image to be detected is added to the uncollected image library. The uncollected image library supports correction of the finger level image. Specifically, when the image to be determined has no bottom level, and / or when the image to be determined has a bottom level but meets a preset condition, the electronic device may add the uncollected image to be detected to the uncollected image library so that the uncollected image to be detected is included in the uncollected image library. In other words, the uncollected image to be detected is used as an uncollected image in the uncollected image library, thereby updating the uncollected image library. Based on this, the uncollected image to be detected can be cached in memory for subsequent base correction of fixed image noise. It should be understood that when the image to be determined has a bottom level but does not meet the preset condition, the uncollected image to be detected may not be added to the uncollected image library, and the uncollected image to be detected may not be used to update the uncollected image library. In embodiments of the present application, after acquiring a sampled image to be detected, a matching sampled image that matches the sampled image to be detected is selected from a sampled image library. This matching sampled image is then used to remove fixed noise from the sampled image to be detected, yielding an image to be determined. This removes fixed signals (i.e., fixed noise) from the sampled image to be detected, eliminating the fixed signals. Furthermore, the fixed signals in the matching sampled image are more similar to those in the sampled image to be detected, improving the accuracy of the image to be determined. Accordingly, when the image to be determined has no bottom level, the sampled image to be detected is added to the sampled image library; and / or, when the image to be determined has a bottom level but meets preset conditions, the sampled image to be detected is added to the sampled image library. The sampled image library supports the correction of finger-level images.As can be seen, the embodiments of the present application can use matching empty-sampled images in the empty-sampled image library that match the empty-sampled image to be detected to perform image processing on the empty-sampled image to be detected, thereby obtaining a highly accurate image to be determined. This allows for bottom-level detection of the empty-sampled image to be detected using the image to be determined, thereby obtaining empty-sampled images with no bottom level or a small bottom-level signal (i.e., empty-sampled images to be detected when the image to be determined has no bottom level and / or when the image to be determined has a bottom level but meets preset conditions). The empty-sampled image library is then updated, and the empty-sampled images in the empty-sampled image library can be empty-sampled images with no bottom level or a small bottom-level signal. Furthermore, the empty-sampled image library supports correction of finger-level images. The empty-sampled images in the empty-sampled image library can then be used to improve the correction effect of finger-level images, thereby reducing the impact of the bottom-level signal on the finger-level image during fixed noise removal, thereby improving the accuracy of finger-level recognition. Based on the above description, the embodiments of the present application also propose a more specific image processing method. Accordingly, the image processing method can be executed by the electronic device mentioned above, or by a chip within the electronic device. For ease of illustration, the following description uses the electronic device executing the image processing method as an example. Referring to FIG. 2 , the image processing method may include the following steps S201-S205:

[0004] 5201, obtain an empty sample image to be detected.

[0005] At step 5202, a matching empty-sampled image that matches the empty-sampled image to be detected is selected from the empty-sampled image library. The matching empty-sampled image is used to remove fixed noise from the empty-sampled image to obtain the image to be determined. In one embodiment, when the matching empty-sampled image is used to remove fixed noise from the empty-sampled image to obtain the image to be determined, the electronic device may perform a difference operation on the empty-sampled image to be detected and the matching empty-sampled image, and the resulting residual image is used as the image to be determined. In other words, the difference operation between the empty-sampled image to be detected and the matching empty-sampled image is used to obtain the residual image, thereby recovering the signal of the empty-sampled image to be detected, and the residual image can be used as the image to be determined. In another embodiment, the electronic device may perform a difference operation on the empty-sampled image to be detected and the matching empty-sampled image to obtain the residual image, and may perform image denoising on the residual image to obtain the image to be determined. Specifically, the residual image after image denoising is used as the image to be determined, thereby achieving the goal of using the matching empty-sampled image to remove fixed noise from the empty-sampled image to obtain the image to be determined. Optionally, the electronic device may employ an image denoising algorithm to perform image denoising on the residual image, thereby making the signal in the image to be determined more distinct. The resulting image to be determined can then be observed to have two types: fingerless and finger-like, with a small number of special features (such as water droplets or dirt). Therefore, the image to be determined has sufficient features for the next step of base-level detection. Optionally, the image denoising algorithm may include, but is not limited to, image smoothing algorithms, frequency-domain Gaussian low-pass filtering algorithms, and Wiener filtering algorithms, among others; this is not limited in this embodiment of the present application. Optionally, the electronic device may include a base correction module, which can then be used by the electronic device to perform step S202 to remove fixed noise from the blank-sampled image to be detected.

[0006] S203: When the image to be determined has no bottom layer, the uncollected image to be detected is added to the uncollected image library; and / or, when the image to be determined has a bottom layer but meets a preset condition, the uncollected image to be detected is added to the uncollected image library. The uncollected image library supports correction of finger-level images. Optionally, the electronic device may further perform image classification on the image to be determined to obtain a bottom layer category of the image to be determined. The bottom layer category may be any of the following: a no-bottom layer category and a bottom layer category. The bottom layer includes at least one of the following: finger-level, water droplets, and dirt (such as dust). Optionally, the image classification may be a binary classification prediction or a multi-class classification prediction, which is not limited in this embodiment of the present application. When the image classification is a multi-class classification prediction, the bottom layer category may be further subdivided into a finger-level category, a water droplet category, and a dirt category. In other words, the bottom layer category may include at least one of the following: finger-level, water droplet, and dirt category. If the bottom-level category is a no-bottom-level category, the image to be determined can be determined to have no bottom-level; if the bottom-level category is a with-bottom-level category, the image to be determined can be determined to have a bottom-level. It should be understood that if the bottom-level category is any of the finger-level category, the water droplet category, and the dirt category among the with-bottom-level categories, the image to be determined to have a bottom-level. In embodiments of the present application, the bottom-level category can be obtained by performing image classification using a bottom-level detection and classification model. The bottom-level detection and classification model can be a convolutional neural network (e.g., a binary classification network or a multi-classification network). Based on this, when performing image classification on the image to be determined to obtain the bottom-level category of the image to be determined, the electronic device can normalize the image to be determined to obtain a normalized image to be determined to meet the input requirements of the convolutional neural network. The electronic device can then call the bottom-level detection and classification model to perform image classification on the normalized image to be determined to obtain the bottom-level category of the image to be determined. In embodiments of the present application, the bottom-level category of the image to be determined can also be used as the bottom-level category of the empty sample image to be detected. Optionally, the embodiment of the present application may use a small convolutional neural network to complete the image classification task. In the feature extraction part, a structure of stacked convolutional layers and pooling layers may be used for feature extraction; in the prediction and regression part, a structure of stacked fully connected layers may be used to obtain the classification prediction result (i.e., the bottom-level category).Optionally, the electronic device may include a convolutional network module. In this case, the electronic device may use the convolutional network module to perform image classification on the image to be determined to obtain the bottom-level category of the image to be determined. Optionally, the electronic device may also obtain an initial bottom-level detection and classification model and a training sampled image set, and determine, from the sampled image library, a training matching sampled image for each training sampled image in the training sampled image set. Training residual images for each training sampled image may then be determined based on the difference calculation results between each training sampled image and the corresponding training matching sampled image. The initial bottom-level detection and classification model may then be invoked to perform image classification on the training residual images of each training sampled image to obtain the bottom-level category of each training residual image. A model loss value may then be calculated based on the bottom-level category of each training residual image and the corresponding label category. Model parameters in the initial bottom-level detection and classification model may then be optimized in a direction that reduces the model loss value to obtain an optimized initial bottom-level detection and classification model. The bottom-level detection and classification model may then be determined based on the optimized initial bottom-level detection and classification model. It should be understood that the electronic device may continue to optimize the optimized initial bottom-level detection and classification model until convergence conditions are met (e.g., the number of iterations reaches a preset iteration threshold, or the model loss value is less than a model loss threshold, etc.), thereby obtaining the bottom-level detection and classification model. Optionally, the preset iteration threshold and model loss threshold may be set based on experience or actual needs, and are not limited in this embodiment of the present application. In other embodiments, the bottom-level detection and classification model may also be a support vector machine classification model, a random forest classification model, a decision tree classification model, a deep learning model, etc.; this is not limited in this embodiment of the present application. Furthermore, if the bottom-level category is a finger-level category among the bottom-level categories, it may be determined that the image to be determined does not meet the preset conditions. If the bottom-level category is a water droplet category or a dirt category among the bottom-level categories, foreign object indication information of the image to be determined may be determined. If the foreign object indication information is less than a preset indication threshold, it may be determined that the image to be determined meets the preset conditions. If the foreign object indication information is greater than or equal to the preset indication threshold, it may be determined that the image to be determined does not meet the preset conditions. The foreign matter includes water droplets and / or dirt.Optionally, when determining foreign object indication information for an image to be determined, the electronic device may perform image recognition on the image to be determined to identify the area where the foreign object (such as water droplets and / or dirt) is located in the image to be determined. Based on the area where the foreign object is located, the electronic device may determine the foreign object area and / or the signal strength in the area where the foreign object is located to determine the foreign object indication information for the image to be determined. Specifically, the foreign object indication information may include at least one of the following: the area of ​​the foreign object (i.e., the area where the foreign object is located) and the signal strength in the area where the foreign object is located. Optionally, the signal strength in the area where the foreign object is located may be the average of the pixel values ​​in the area where the foreign object is located, or the maximum pixel value in the area where the foreign object is located, etc.; this is not limited in this embodiment of the present application. Optionally, the foreign object indication information may include the area of ​​the foreign object and / or the signal strength in the area where the foreign object is located. Accordingly, the preset indication threshold may include a preset area threshold and / or a preset strength threshold. Optionally, if the foreign object indication information includes the foreign object area or the signal strength in the area where the foreign object is located, then if the foreign object area is less than a preset area threshold, or the signal strength in the area where the foreign object is located is less than a preset strength threshold, the foreign object indication information may be determined to be less than the preset indication threshold. If the foreign object area is greater than or equal to the preset area threshold, or the signal strength in the area where the foreign object is located is greater than or equal to the preset strength threshold, the foreign object indication information may be determined to be greater than or equal to the preset indication threshold. Accordingly, if the foreign object indication information includes both the foreign object area and the signal strength in the area where the foreign object is located, then if the foreign object area is less than the preset area threshold and the signal strength in the area where the foreign object is located is less than the preset strength threshold, the foreign object indication information may be determined to be less than the preset indication threshold. If the foreign object area is greater than or equal to the preset area threshold, or the signal strength in the area where the foreign object is located is greater than or equal to the preset strength threshold, the foreign object indication information may be determined to be greater than or equal to the preset indication threshold. Optionally, both the preset area threshold and the preset strength threshold may be set based on experience or actual needs, and this is not limited in this embodiment of the present application. In other words, the preset indication threshold may be set based on experience or actual needs.Optionally, when updating the air-collected image library using the air-collected image to be detected (i.e., when the air-collected image to be detected is added to the air-collected image library), the electronic device may also add the image indication dataset of the air-collected image to be detected to the air-collected image library. If the air-collected image to be detected has not been added to the air-collected image library, the image indication dataset of the air-collected image to be detected may not be added to the air-collected image library. For example, features such as the acquisition frequency and flight time of the air-collected image to be detected may also be stored in the air-collected image library for subsequent image matching and, therefore, for subsequent base correction.

[0007] S204: When a need to remove an uncollected image library is detected, the uncollected images to be removed are determined from the uncollected image library. Optionally, the electronic device may determine that a need to remove an uncollected image library has been detected when it detects that the uncollected image library has reached its upper limit. Alternatively, the uncollected image library may further include image gradients for each uncollected image, and may determine that a need to remove an uncollected image library has been detected when it detects that the number of uncollected images in the uncollected image library exceeds a preset storage threshold, and that an uncollected image in the uncollected image library has an image gradient greater than a preset gradient threshold. This is not limited in this embodiment of the present application. Optionally, the upper limit of the library capacity, the preset storage threshold, and the preset gradient threshold may all be set based on experience or actual needs, and are not limited in this embodiment of the present application. In one specific implementation, the electronic device may determine the time each uncollected image in the uncollected image library was added, and select the uncollected image with the earliest addition time in the uncollected image library as the uncollected image to be removed. Optionally, the uncollected image library may store the addition time of each uncollected image, thereby determining the addition time of each uncollected image. Alternatively, each uncollected image in the uncollected image library may be updated in a queue in the order of addition time. In this case, the first uncollected image in the uncollected image library may be located before the second uncollected image, and the addition time of the first uncollected image may be earlier than the addition time of the second uncollected image. In other words, the addition time of each uncollected image may be determined based on the order of the uncollected images in the uncollected image library, and the addition time of each uncollected image may be expressed in order. For example, the addition time of the uncollected image in the first position may be 1, where a smaller value indicates an earlier addition time, or an uncollected image located closer to the front of the image indicates an earlier addition time, and so on. Based on this, the electronic device may eliminate excess cache bases (i.e., uncollected images to be eliminated) according to the first-in-first-out principle.In another specific implementation, the empty-sampled image library can also be used to store image gradients associated with each empty-sampled image. Based on this, when an empty-sampled image to be detected is added to the empty-sampled image library, the image gradient associated with the empty-sampled image to be detected can be determined using the image to be determined, and the image gradient associated with the empty-sampled image to be detected can be added to the empty-sampled image library. When the empty-sampled image to be detected has not been added to the empty-sampled image library, the image gradient associated with the empty-sampled image to be detected is not calculated. In this case, the electronic device can determine the image gradient associated with each empty-sampled image and, based on the image gradient associated with each empty-sampled image, select an empty-sampled image to be eliminated from the empty-sampled image library. The image gradient associated with the empty-sampled image to be eliminated is greater than the image gradient associated with empty-sampled images in the empty-sampled image library other than the empty-sampled image to be eliminated, or the image gradient associated with the empty-sampled image to be eliminated is null. The image gradient associated with the initial empty-sampled image in the empty-sampled image library is null. It should be understood that the blank image library may initially contain at least one initial blank image. That is, at least one initial blank image may be directly added to the blank image library for subsequent image matching. Optionally, the at least one initial blank image may be set based on experience, actual needs, or the first Q acquired blank images, where Q is a positive integer, etc., and this is not limited in the present embodiment. Optionally, if the blank image library contains blank images with a null image gradient, a blank image with a null image gradient may be first identified from the blank image library and used as the blank image to be removed. For example, the blank image with a null image gradient that was added the earliest may be identified, or a random blank image with a null image gradient may be selected. If no blank image with a null image gradient exists in the blank image library, the blank image with the largest image gradient in the blank image library may be used as the blank image to be removed. Optionally, the electronic device may also select, from the empty-sampled image library, empty-sampled images whose image gradients are greater than a preset gradient threshold as empty-sampled images to be removed, thereby preemptively removing empty-sampled images with large image gradients. Based on this, embodiments of the present application can remove empty-sampled images with potentially low-level features that are difficult for the low-level detection and classification model to detect from the empty-sampled image library, thereby preventing these empty-sampled images from affecting the correction process and further improving the correction effect.Specifically, when using the image to be determined to determine the image gradient under the blank-sampled image to be detected, the image to be determined can be used to determine the gradient calculation image grayscale matrix, and then the gradient calculation image grayscale matrix can be processed using a target operator to obtain an image gradient matrix. The image gradient matrix can then be used to determine the image gradient under the blank-sampled image to be detected. For example, the image gradient matrix can be used as the image gradient under the blank-sampled image to be detected, or the mean of the elements in the image gradient matrix can be used as the image gradient under the blank-sampled image to be detected. Optionally, the target operator can be a Sobel operator (a type of pixel image edge detection operator) or a Schan operator (another type of pixel image edge detection operator), etc., which is not limited in this embodiment of the present application. Optionally, the gradient-calculated image grayscale matrix may be the image grayscale matrix of the image to be determined, or may be the normalized result of the image grayscale matrix of the image to be determined, etc., which is not limited in the embodiments of the present application. Accordingly, the image gradient can be used to indicate the degree of change in pixel values ​​in the corresponding gradient-calculated image grayscale matrix. The greater the image gradient, the greater the degree of change, and the more likely the gradient-calculated image grayscale matrix contains noise, that is, the more likely the image to be determined contains noise. Therefore, the image gradient can be used to indicate the likelihood that the corresponding blank image contains a bottom level. For example, taking the Sobel operator as an example for further explanation, the electronic device can use the Sobel operator to process the gradient-calculated image grayscale matrix to obtain the image gradient matrix. Specifically, the electronic device can use Formula 2.1 to calculate the image gradient matrix:

[0008] G = eye + G; Formula 2.1, where G can be the image gradient matrix, Gx can be the horizontal gradient, and Gy can be the vertical gradient. Optionally, the electronic device can use Formula 2.2 to calculate Gx: Calculate the convolution kernel of the horizontal gradient. Accordingly, the electronic device can use formula 2.3 to calculate Gy:

[0009] S205: Remove the empty-sampled images to be removed from the empty-sampled image library. It should be understood that the electronic device can remove additional information (such as image gradient, flight time, and frequency) of the empty-sampled images to be removed from the empty-sampled image library, thereby removing the image indication dataset of the empty-sampled images to be removed from the empty-sampled image library. Optionally, the electronic device can include an empty-sampled image library update module (i.e., a base library update module). In this case, the electronic device can use the empty-sampled image library update module to add the empty-sampled images to be detected to the empty-sampled image library, remove the empty-sampled images to be removed from the empty-sampled image library, and so on. Exemplarily, as shown in FIG3 , the electronic device may execute step S301 to construct a sampled image library, and may execute step S302 to obtain a sampled image to be detected; correspondingly, step S303 may be executed to obtain a matching sampled image (i.e., obtain a matching sampled image that matches the sampled image to be detected from the sampled image library) through a base correction module, and use the matching sampled image to remove fixed noise from the sampled image to be detected to obtain an image to be determined; and, step S304 may be executed to perform image classification on the image to be determined through a convolutional network module to obtain a bottom-level category of the image to be determined. Based on this, step S305 may be executed to determine whether the uncollected image to be detected includes foreign matter, that is, whether the uncollected image to be detected includes foreign matter may be determined based on the bottom-level category of the image to be determined. When it is determined that the uncollected image to be detected does not include foreign matter (i.e., the image to be determined has no bottom level and / or the image to be determined has a bottom level but satisfies a preset condition), step S306 may be executed to add the uncollected image to be detected to the uncollected image library through the uncollected image library update module. When it is determined that the uncollected image to be detected includes foreign matter (i.e., the image to be determined has a bottom level but does not satisfy the preset condition), the uncollected image to be detected may not be added to the uncollected image library, that is, the uncollected image library update module may not be used at this time, that is, step S306 may not be executed. Optionally, the electronic device may further use the uncollected image library to remove fixed noise from the finger-level image, thereby obtaining a corrected finger-level image, so as to perform finger-level recognition on the corrected finger-level image.Specifically, the electronic device may obtain a finger-level image to be identified, which includes the user's finger-level information. Then, the electronic device may select a sampled image from a sampled image library for use in correcting the finger-level image to be identified. A difference operation is performed between the finger-level image to be identified and the sampled image to remove fixed noise from the finger-level image to obtain a corrected finger-level image to be identified. Based on this difference, finger-level recognition may be performed on the corrected finger-level image to obtain a finger-level recognition result. Optionally, the electronic device may select a sampled image from the sampled image library for use in correcting the finger-level image to be identified, using the aforementioned matching sampled image determination method, thereby further improving the accuracy of the finger-level recognition result. It should be noted that the ultrasonic base image (i.e., the air-sampled image) can effectively capture the fixed noise signal collected by the ultrasonic sensor (i.e., the ultrasonic finger-level sensor). Subtracting the image from the ultrasonic finger-level image can remove the interference of this fixed noise signal. However, the ultrasonic base image may also capture non-stationary background signals, such as finger-levels, water droplets, and dirt (e.g., dust). Subtracting these non-stationary signals from the ultrasonic finger-level image may interfere with the finger-level signal. Therefore, embodiments of the present application can detect ultrasonic base images with background signals and construct an air-sampled image library using air-sampled images with no background signals and / or low background signals. The air-sampled images in the air-sampled image library can then be used to remove fixed noise from the finger-level image, effectively preventing interference from air-sampled images with large background signals on the finger-level signal (i.e., the finger-level image). After acquiring an undetected image, embodiments of the present application select a matching undetected image from an undetected image library. Using this matching image, fixed noise removal is performed on the undetected image to obtain an image to be determined. Therefore, if the image to be determined does not have a bottom level, the undetected image can be added to the undetected image library. And / or, if the image to be determined does have a bottom level but meets preset conditions, the undetected image can be added to the undetected image library, which supports the correction of finger-level images. Furthermore, if a need to remove an undetected image from the library is detected, the undetected image to be removed can be determined from the library and removed from the library.As can be seen, the embodiments of the present application propose a bottom-level detection solution for ultrasonic finger-level sensors in blank-sampled images. Through methods such as framework design and deep learning network detection, the blank-sampled images to be detected can be processed to identify blank-sampled images with a bottom-level, and a base-level-free blank-sampled image library can be constructed. This prevents base correction in the ultrasonic finger-level image processing process from generating abnormal bottom-levels. This effectively reduces the occurrence of abnormal signals in subsequent algorithm processes that can lead to recognition errors, thereby effectively improving the overall performance of ultrasonic finger-level recognition. Based on the description of the relevant embodiments of the above-mentioned image processing method, the embodiments of the present application also propose an image processing device. The image processing device can be a computer program (including program code) running in an electronic device. Optionally, the electronic device can include a chip, and the image processing device can also be a computer program running in the chip included in the electronic device. As shown in Figure 4, the image processing device can include an acquisition unit 401 and a processing unit 402. The image processing device can execute the image processing method shown in Figure 1 or Figure 2, that is, the image processing device can run the above units: an acquisition unit 401, used to acquire an empty-sampled image to be detected; a processing unit 402, used to select a matching empty-sampled image that matches the empty-sampled image to be detected from the empty-sampled image library, and use the matching empty-sampled image to remove fixed noise from the empty-sampled image to be detected to obtain an image to be determined; the processing unit 402 is also used to add the empty-sampled image to be detected to the empty-sampled image library when the image to be determined has no bottom level; and / or, when the image to be determined has a bottom level but meets a preset condition, add the empty-sampled image to be detected to the empty-sampled image library, and the empty-sampled image library supports correction of finger-level images.In one embodiment, when selecting a matching air-sampled image that matches the air-sampled image to be detected from the air-sampled image library, the processing unit 402 may be specifically configured to: determine an image indication dataset of the air-sampled image to be detected, select M target air-sampled images from the air-sampled image library, and determine an image indication dataset of each target air-sampled image in the M target air-sampled images, where the image indication dataset of one image includes at least one of the following: an image grayscale matrix of the corresponding image, an acquisition frequency of the corresponding image, an acquisition temperature of the corresponding image, an acquisition time of the corresponding image, and a flight time of the corresponding image, where M is a positive integer; the acquisition frequency of the corresponding image is a signal transmission frequency of an ultrasonic signal that generates the image, and the flight time of the corresponding image is a signal flight time of the ultrasonic signal that generates the image; and respectively calculate similarities between the image indication dataset of the air-sampled image to be detected and the image indication datasets of each target air-sampled image, where the similarities between the image indication dataset of the air-sampled image to be detected and the image indication dataset of one target air-sampled image are used to indicate the similarity between the air-sampled image to be detected and the corresponding target air-sampled image. Based on the similarity between the unmined image to be detected and the target unmined images, the target unmined image with the greatest similarity to the unmined image to be detected among the M target unmined images is used as the matching unmined image to be matched with the unmined image to be detected.In another embodiment, when the processing unit 402 respectively calculates the similarity between the image indication data set of the unmined image to be detected and the image indication data set of each target unmined image, it can be specifically used to: determine, for any target unmined image among the M target unmined images, and for any image indication data in the image indication data set of the any target unmined image, a similarity calculation method corresponding to any image indication data of the any target unmined image; according to the similarity calculation method, based on any image indication data of the any target unmined image and the image indication data of the unmined image to be detected, calculate the similarity under any image indication data of the any target unmined image to obtain the similarity under each image indication data of the any target unmined image; the similarity under any image indication data of the any target unmined image refers to: the similarity between any image indication data of the any target unmined image and the image indication data of the unmined image to be detected, and the type of the image indication data of the unmined image to be detected is the same as the type of the image indication data of the any target unmined image; A weighted summation is performed on the similarities under each image indication data of any target uncollected image to obtain the similarity between the image indication data set of the uncollected image to be detected and the image indication data set of any target uncollected image. In another embodiment, when selecting M target uncollected images from the uncollected image library, the processing unit 402 may be specifically configured to: add each uncollected image in the uncollected image library to the M target uncollected images, so that the M target uncollected images include each uncollected image in the uncollected image library; or, for any uncollected image in the uncollected image library, determine at least one image indication data from the image indication data set of any uncollected image; and if the difference between each image indication data in the at least one image indication data and the corresponding image indication data of the uncollected image to be detected is less than a corresponding preset difference threshold, then add the any uncollected image to the M target uncollected images to obtain the M target uncollected images.In another embodiment, the processing unit 402 may further be used to: perform image classification on the image to be determined to obtain a bottom-level category of the image to be determined, where the bottom-level category is any one of the following: a category without a bottom level and a category with a bottom level; wherein the bottom level includes at least one of the following: fingers, water droplets, and dirt; if the bottom-level category is the category without a bottom level, determining that the image to be determined has no bottom level; if the bottom-level category is the category with a bottom level, determining that the image to be determined has a bottom level. In another embodiment, the bottom-level category includes at least one of the following: a finger-level category, a water droplet category, and a dirt category; and the processing unit 402 may further be configured to: if the bottom-level category is the finger-level category among the bottom-level categories, determine that the image to be determined does not meet the preset condition; if the bottom-level category is the water droplet category or the dirt category among the bottom-level categories, determine foreign object indication information for the image to be determined; if the foreign object indication information is less than a preset indication threshold, determine that the image to be determined meets the preset condition; and if the foreign object indication information is greater than or equal to the preset indication threshold, determine that the image to be determined does not meet the preset condition. In another embodiment, the foreign object indication information includes at least one of the following: the area of ​​the foreign object and the signal strength of the area where the foreign object is located; wherein the foreign object includes water droplets and / or dirt. In another embodiment, the bottom-level category is obtained by performing image classification using a bottom-level detection and classification model, wherein the bottom-level detection and classification model is a convolutional neural network. When performing image classification on the image to be determined to obtain the bottom-level category of the image to be determined, the processing unit 402 may be specifically configured to: normalize the image to be determined to obtain a normalized image to be determined; invoke the bottom-level detection and classification model to perform image classification on the normalized image to be determined to obtain the bottom-level category of the image to be determined. In another embodiment, the processing unit 402 may also be configured to: upon detecting a need to remove the empty-collected image library, determine the empty-collected image to be removed from the empty-collected image library; and remove the empty-collected image to be removed from the empty-collected image library.In another embodiment, the processing unit 402 may further be configured to: when the uncollected image to be detected is added to the uncollected image library, use the image to be determined to determine the image gradient under the uncollected image to be detected, and add the image gradient under the uncollected image to the uncollected image library; when the uncollected image to be detected is not added to the uncollected image library, not calculate the image gradient under the uncollected image to be detected. In another embodiment, when the processing unit 402 determines the empty-sampled images to be eliminated from the empty-sampled image library, it can be specifically used to: determine the addition time of each empty-sampled image in the empty-sampled image library, and use the empty-sampled image with the earliest addition time in the empty-sampled image library as the empty-sampled image to be eliminated; or, determine the image gradient under each empty-sampled image, and based on the image gradient under each empty-sampled image, select the empty-sampled image to be eliminated from the empty-sampled image library, the image gradient under the empty-sampled image to be eliminated is greater than the image gradient under the empty-sampled images other than the empty-sampled image to be eliminated in the empty-sampled image library, or the image gradient under the empty-sampled image to be eliminated is empty; wherein, the image gradient under the initial empty-sampled image in the empty-sampled image library is empty. In another embodiment, the processing unit 402 may be further configured to: when the uncollected image to be detected is added to the uncollected image library, add the image indication dataset of the uncollected image to the uncollected image library; and when the uncollected image to be detected is not added to the uncollected image library, not add the image indication dataset of the uncollected image to the uncollected image library. In another embodiment, when the processing unit 402 uses the matching uncollected image to perform fixed noise removal on the uncollected image to be detected to obtain the image to be determined, the processing unit 402 may be specifically configured to: perform a difference operation on the uncollected image to be detected and the matching uncollected image, and obtain a residual image as the image to be determined; or perform a difference operation on the uncollected image to be detected and the matching uncollected image to obtain a residual image, and perform image noise reduction on the residual image to obtain the image to be determined.In another embodiment, the processing unit 402 may further be configured to: obtain a finger-level image to be identified, the finger-level image to be identified including user finger-level information; select a sampled image from the sampled image library for use in correcting the finger-level image to be identified, and perform a difference operation between the finger-level image to be identified and the sampled image to remove fixed noise from the finger-level image to be identified, thereby obtaining a corrected finger-level image to be identified; and perform finger-level recognition on the corrected finger-level image to be identified, thereby obtaining a finger-level recognition result. According to one embodiment of the present application, each step involved in the method shown in FIG. 1 or FIG. 2 may be performed by each unit in the image processing device shown in FIG. 4 . For example, step S101 shown in FIG. 1 may be performed by the acquisition unit 401 shown in FIG. 4 , and steps S102 and S103 may be performed by the processing unit 402 shown in FIG. 4 . For another example, step S201 shown in FIG2 can be performed by the acquisition unit 401 shown in FIG4 , and steps S202-S205 can all be performed by the processing unit 402 shown in FIG4 , and so on. According to another embodiment of the present application, each unit in the image processing device shown in FIG4 can be individually or entirely combined into one or more other units, or one or more of the units can be further divided into multiple functionally smaller units. This can achieve the same operation without affecting the technical effects of the embodiments of the present application. The above-mentioned units are divided based on logical functions. In actual applications, the functions of one unit can be implemented by multiple units, or the functions of multiple units can be implemented by one unit. In other embodiments of the present application, any image processing device may also include other units. In actual applications, these functions can also be implemented with the assistance of other units, or can be implemented collaboratively by multiple units. According to another embodiment of the present application, an image processing device as shown in FIG. 4 can be constructed, and the image processing method of the embodiment of the present application can be implemented by running a computer program (including program code) capable of executing the steps involved in the corresponding method shown in FIG. 1 or FIG. 2 on a general electronic device such as a computer that includes processing elements and storage elements such as a central processing unit (CPU), a random access memory (RAM), and a read-only memory (ROM).The computer program can be recorded on, for example, a computer storage medium and loaded into the electronic device via the computer storage medium and executed therein. In an embodiment of the present application, after acquiring the undetected image, a matching undetected image that matches the undetected image can be selected from an undetected image library. The matching undetected image can then be used to remove fixed noise from the undetected image to obtain an image to be determined. This removes fixed signals (i.e., fixed noise) from the undetected image, eliminating the fixed signals. Furthermore, the fixed signals in the matching undetected image are more similar to those in the undetected image, thereby improving the accuracy of the image to be determined. Accordingly, when the image to be determined has no bottom level, the undetected image is added to the undetected image library; and / or, when the image to be determined has a bottom level but meets preset conditions, the undetected image is added to the undetected image library. The undetected image library supports the correction of finger-level images. As can be seen, embodiments of the present application can process the undetected image using matching undetected images in the undetected image library to obtain a more accurate image to be determined. This allows for bottom-level detection of the undetected image using the image to be determined, thereby obtaining undetected images with no bottom-level or low bottom-level signals (i.e., undetected images with no bottom-level or low bottom-level signals when the image to be determined has no bottom-level and / or has a bottom-level but meets preset conditions). The undetected image library is then updated, and the undetected images in the undetected image library can be undetected images with no bottom-level or low bottom-level signals. Furthermore, the undetected image library supports correction of finger-level images. The undetected images in the undetected image library can be used to improve finger-level image correction, thereby reducing the impact of the bottom-level signal on the finger-level image during fixed noise removal, thereby improving finger-level recognition accuracy. Based on the description of the above method and apparatus embodiments, exemplary embodiments of the present application further provide an electronic device comprising: one or more processors; and one or more memories communicatively connected to the one or more processors. The one or more memories store a computer program that can be executed by the one or more processors. When the computer program is executed by the one or more processors, it is used to enable the electronic device to perform the method according to the embodiment of the present application.The exemplary embodiments of the present application also provide a chip, located within an electronic device, configured to execute a method according to an embodiment of the present application. The exemplary embodiments of the present application also provide a non-transitory computer-readable storage medium storing a computer program, wherein the computer program, when executed by a computer processor, causes the computer to execute the method according to an embodiment of the present application. The exemplary embodiments of the present application also provide a computer program product, including the computer program, wherein the computer program, when executed by a computer processor, causes the computer to execute the method according to an embodiment of the present application. Referring to FIG. 5 , a block diagram of an electronic device 500 that can serve as a server or client in the present application will now be described. This is an example of a hardware device applicable to various aspects of the present application. The term "electronic device" is intended to refer to various forms of digital electronic computing devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The term "electronic device" may also refer to various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components, their connections and relationships, and their functions shown herein are merely examples and are not intended to limit the implementation of the present application as described and / or claimed herein. As shown in FIG5 , electronic device 500 includes a computing unit 501, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. RAM 503 may also store various programs and data required for the operation of electronic device 500. Computing unit 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to bus 504. Multiple components in electronic device 500 are connected to I / O interface 505, including an input unit 506, an output unit 507, a storage unit 508, and a communication unit 509.Input unit 506 can be any type of device capable of inputting information into electronic device 500. Input unit 506 can receive input numeric or character information and generate key signal inputs related to user settings and / or function control of the electronic device. Output unit 507 can be any type of device capable of presenting information, and may include, but is not limited to, a display, a speaker, a video / audio output terminal, a vibrator, and / or a printer. Storage unit 508 may include, but is not limited to, a magnetic disk or an optical disk. Communication unit 509 allows electronic device 500 to exchange information / data with other devices via computer networks such as the Internet and / or various telecommunication networks, and may include, but is not limited to, a modem, a network card, an infrared communication device, a wireless communication transceiver, and / or a chipset, such as a Bluetooth™ device, a WiFi device, a WiMax device, a cellular communication device, and / or the like. Computing unit 501 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 501 performs the various methods and processes described above. For example, in some embodiments, the image processing method may be implemented as a computer software program that is tangibly contained in a machine-readable medium, such as the storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 500 via the ROM 502 and / or the communication unit 509. In some embodiments, the computing unit 501 may be configured to perform the image processing method in any other appropriate manner (e.g., via firmware). The program code for implementing the methods of the present application may be written in any combination of one or more programming languages. Such program code may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the program code is executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented.The program code may execute entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server. In the context of this application, a machine-readable medium may be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and / or device (e.g., a magnetic disk, an optical disk, a memory, a programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including machine-readable media that receive machine instructions as machine-readable signals. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor. To provide user interaction, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), a computing system that includes middleware components (e.g., an application server), a computing system that includes frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected via any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet. A computer system can include clients and servers. The clients and servers are generally remote from each other and typically interact via a communication network. The client-server relationship arises through computer programs running on the respective computers and establishing a client-server relationship with each other. Furthermore, it should be understood that the above disclosures are merely preferred embodiments of the present application and are not intended to limit the scope of the present application. Therefore, equivalent variations based on the claims of the present application are still within the scope of the present application.

Claims

Claims 1. An image processing method, characterized in that: include: Acquire an empty-sampled image to be detected; select a matching empty-sampled image that matches the empty-sampled image to be detected from an empty-sampled image library, and use the matching empty-sampled image to remove fixed noise from the empty-sampled image to be detected to obtain an image to be determined; when the image to be determined has no bottom level, add the empty-sampled image to be detected to the empty-sampled image library; and / or, when the image to be determined has a bottom level but meets a preset condition, add the empty-sampled image to be detected to the empty-sampled image library, and the empty-sampled image library supports correction of finger-level images.

2. The method according to claim 1, characterized in that The selecting, from the air-collected image library, a matching air-collected image that matches the air-collected image to be detected includes: determining an image indication dataset of the air-collected image to be detected, selecting M target air-collected images from the air-collected image library, and determining an image indication dataset of each target air-collected image in the M target air-collected images, wherein the image indication dataset of one image includes at least one of the following: an image grayscale matrix of a corresponding image, an acquisition frequency of the corresponding image, an acquisition temperature of the corresponding image, an acquisition time of the corresponding image, and a flight time of the corresponding image, where M is a positive integer; the acquisition frequency of the corresponding image is a signal transmission frequency of an ultrasonic signal that generates the image, and the flight time of the corresponding image is a signal flight time of the ultrasonic signal that generates the image; respectively calculating similarities between the image indication dataset of the air-collected image to be detected and the image indication datasets of each target air-collected image, wherein the similarity between the image indication dataset of the air-collected image to be detected and an image indication dataset of a target air-collected image is used to indicate the similarity between the air-collected image to be detected and the corresponding target air-collected image; Based on the similarity between the unmined image to be detected and the target unmined images, the target unmined image with the greatest similarity to the unmined image to be detected among the M target unmined images is used as the matching unmined image to be matched with the unmined image to be detected.

3. The method according to claim 2, wherein The similarity between the image indication dataset of the to-be-detected empty-sampled image and the image indication dataset of each target empty-sampled image is calculated respectively, including: for any target empty-sampled image among the M target empty-sampled images, and for any target empty-sampled image 28 any image indication data in an image indication data set of a target un-mined image, determine a similarity calculation method corresponding to any image indication data of the any target un-mined image; according to the similarity calculation method, based on any image indication data of the any target un-mined image and the image indication data of the un-mined image to be detected, calculate the similarity under any image indication data of the any target un-mined image to obtain the similarity under each image indication data of the any target un-mined image; the similarity under any image indication data of the any target un-mined image refers to: the similarity between any image indication data of the any target un-mined image and the image indication data of the un-mined image to be detected, the type of the image indication data of the un-mined image to be detected is the same as the type of any image indication data of the any target un-mined image; weighted summation is performed on the similarities under each image indication data of the any target un-mined image to obtain the similarity between the image indication data set of the un-mined image to be detected and the image indication data set of the any target un-mined image.

4. The method according to claim 2, wherein The selecting M target empty-sampled images from the empty-sampled image library includes: adding each empty-sampled image in the empty-sampled image library to the M target empty-sampled images, so that the M target empty-sampled images include each empty-sampled image in the empty-sampled image library; or, for any empty-sampled image in the empty-sampled image library, determining at least one image indication data from the image indication data set of any empty-sampled image; if the difference between each image indication data in the at least one image indication data and the corresponding image indication data of the empty-sampled image to be detected is less than the corresponding preset difference threshold, then adding the any empty-sampled image to the M target empty-sampled images to obtain the M target empty-sampled images.

5. The method according to any one of claims 1 to 4, characterized in that The method further includes: performing image classification on the image to be determined to obtain a bottom-level category of the image to be determined, where the bottom-level category is any one of the following: a category without a bottom level and a category with a bottom level; wherein the bottom level includes at least one of the following: fingers, water droplets, and dirt; if the bottom-level category is the category without a bottom level, determining that the image to be determined has no bottom level; if the bottom-level category is the category with a bottom level, determining that the image to be determined has a bottom level.

6. The method according to claim 5, wherein The bottom-level categories include the following At least one of: a finger-level category, a water droplet category, and a dirt category; The method also includes: if the bottom-level category is a finger-level category among the bottom-level categories, determining that the image to be determined does not meet the preset conditions; if the bottom-level category is a water droplet category or a dirt category among the bottom-level categories, determining the foreign matter indication information of the image to be determined, if the foreign matter indication information is less than a preset indication threshold, determining that the image to be determined meets the preset conditions; if the foreign matter indication information is greater than or equal to the preset indication threshold, determining that the image to be determined does not meet the preset conditions.

7. The method according to claim 6, wherein The foreign object indication information includes at least one of the following: the area of the foreign object and the signal strength of the area where the foreign object is located; wherein the foreign object includes water droplets and / or dirt.

8. The method according to claim 5, wherein The bottom-level category is obtained by performing image classification using a bottom-level detection and classification model, where the bottom-level detection and classification model is a convolutional neural network. Classifying the image to be determined to obtain the bottom-level category of the image to be determined includes: normalizing the image to be determined to obtain a normalized image to be determined; and calling the bottom-level detection and classification model to perform image classification on the normalized image to be determined to obtain the bottom-level category of the image to be determined.

9. The method according to any one of claims 1 to 4, wherein The method further includes: when a need to remove the empty-sampled image library is detected, determining the empty-sampled images to be removed from the empty-sampled image library; and removing the empty-sampled images to be removed from the empty-sampled image library.

10. The method according to claim 9, characterized in that The method further includes: when the uncollected image to be detected is added to the uncollected image library, using the image to be determined, determining the image gradient under the uncollected image to be detected, and adding the image gradient under the uncollected image to the uncollected image library; when the uncollected image to be detected is not added to the uncollected image library, not calculating the image gradient under the uncollected image to be detected.

11. The method according to claim 10, characterized in that Determining the empty-sampled images to be eliminated from the empty-sampled image library includes: determining the adding time of each empty-sampled image in the empty-sampled image library, and taking the empty-sampled image with the earliest adding time in the empty-sampled image library as the empty-sampled image to be eliminated; or, determining the image gradient under each empty-sampled image, and selecting the empty-sampled image to be eliminated from the empty-sampled image library based on the image gradient under each empty-sampled image, the image gradient under the empty-sampled image to be eliminated is greater than the image gradient under the empty-sampled images other than the empty-sampled image to be eliminated in the empty-sampled image library, or the image gradient under the empty-sampled image to be eliminated is empty; wherein, the image gradient under the initial empty-sampled image in the empty-sampled image library is empty.

12. The method according to any one of claims 1 to 4, characterized in that The method also includes: when the empty-sampled image to be detected is added to the empty-sampled image library, adding the image indication dataset of the empty-sampled image to be detected to the empty-sampled image library; when the empty-sampled image to be detected is not added to the empty-sampled image library, not adding the image indication dataset of the empty-sampled image to be detected to the empty-sampled image library.

13. The method according to any one of claims 1 to 4, characterized in that The method of using the matching empty-sampled image to remove fixed noise from the empty-sampled image to be detected to obtain the image to be determined includes: performing a difference operation on the empty-sampled image to be detected and the matching empty-sampled image, and obtaining a residual image as the image to be determined; or performing a difference operation on the empty-sampled image to be detected and the matching empty-sampled image to obtain a residual image, and performing image denoising processing on the residual image to obtain the image to be determined.

14. The method according to any one of claims 1 to 4, characterized in that The method further comprises: obtaining a finger level image to be identified, the finger level image to be identified including user finger level information; selecting a blank image for correcting the finger level image to be identified from the blank image library, and performing a difference operation on the finger level image to be identified and the blank image to correct the finger level image to be identified. The fixed noise of the finger-level image to be identified is removed to obtain a corrected finger-level image to be identified; and finger-level recognition is performed on the corrected finger-level image to be identified to obtain a finger-level recognition result.

15. An image processing device, characterized in that: The device includes: an acquisition unit, used to acquire an empty-sampled image to be detected; a processing unit, used to select a matching empty-sampled image that matches the empty-sampled image to be detected from an empty-sampled image library, and use the matching empty-sampled image to remove fixed noise from the empty-sampled image to be detected to obtain an image to be determined; the processing unit is also used to add the empty-sampled image to be detected to the empty-sampled image library when the image to be determined has no bottom level; and / or, when the image to be determined has a bottom level but meets a preset condition, add the empty-sampled image to be detected to the empty-sampled image library, and the empty-sampled image library supports correction of finger-level images.

16. A chip, characterized in that: The chip is located in an electronic device, and the chip is configured to execute the method according to any one of claims 1 to 14.

17. An electronic device, comprising: One or more processors; and one or more memories storing a program, wherein the program comprises instructions that, when executed by the one or more processors, cause the one or more processors to perform the method according to any one of claims 1 to 14.

18. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause a computer to execute the method according to any one of claims 1-14. 32

Images