CN114792289A - Image pooling method, apparatus, terminal device and computer-readable storage medium - Google Patents

Abstract

Embodiments of the present application provide an image pooling method, apparatus, terminal device, and computer-readable storage medium. The method includes: acquiring an image to be processed; performing a first pooling process on the image to be processed to obtain a reference image; The pixels of the processed image are given corresponding weights; according to the weights, the second pooling processing is performed on the to-be-processed image and the reference image to obtain the target image. In this embodiment of the present application, the pixels at different positions of the image to be processed are given weights according to the weight formula, and the image to be processed and the reference image are subjected to a second pooling process according to the weights to obtain the final target image, thereby retaining more image details and making more It conforms to the subjective feeling of human vision.

Description

Image pooling method, apparatus, terminal device, and computer-readable storage medium

technical field

The present application belongs to the technical field of image processing, and in particular, relates to an image pooling method, apparatus, terminal device and computer-readable storage medium.

Background technique

With the continuous development of AI technology, image-based neural network technology has also entered a stage of rapid development. In today's world, the image resolution is getting higher and higher, and the pooling layer has become a common necessary network layer to reduce the resolution during the training process of the neural network. It has significant characteristics such as increasing the receptive field, not deforming in translation, and reducing the difficulty of neural network optimization. It is an indispensable part of the existing neural network and has great significance.

The existing pooling layers include average pooling, maximum pooling, and random pooling. Although these pooling methods are common and easy to use, they have their own characteristics in neural networks, but these existing methods blur the edges and lose a lot of details. , resulting in poor visual subjective experience, which cannot meet the human vision's attention to semantic information and edge details.

SUMMARY OF THE INVENTION

The embodiments of the present application provide an image pooling method, apparatus, terminal device, and computer-readable storage medium, which are used to solve the problem of insufficient image details in the existing pooling technology.

In a first aspect, an embodiment of the present application provides an image pooling method, including:

Get the image to be processed;

Perform a first pooling process on the image to be processed to obtain a reference image;

Give corresponding weights to the pixels of the image to be processed according to the reference image;

According to the weight, the second pooling process is performed on the to-be-processed image and the reference image to obtain the target image.

In a second aspect, an embodiment of the present application provides an image processing apparatus, including:

The acquisition module is used to acquire the image to be processed;

The first pooling module is used to perform first pooling processing on the image to be processed to obtain a reference image;

The weight module is used to assign corresponding weights to the pixels of the image to be processed according to the reference image;

The second pooling module is used to perform a second pooling process on the image to be processed and the reference image according to the weight to obtain the target image.

In a third aspect, an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, any one of the foregoing first aspects is implemented The steps of the image pooling method for the item.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the image pooling method according to any one of the foregoing first aspects is implemented. step.

In a fifth aspect, an embodiment of the present application provides a computer program product that, when the computer program product runs on a terminal device, causes the terminal device to perform the steps of the image pooling method in any one of the above-mentioned first aspects.

It can be understood that, for the beneficial effects of the foregoing second aspect to the fifth aspect, reference may be made to the relevant description in the foregoing first aspect, which will not be repeated here.

Compared with the prior art, the beneficial effects of the embodiments of the present application are as follows: the embodiments of the present application provide an image pooling method, apparatus, terminal device and computer-readable storage medium. The method includes: acquiring an image to be processed; The image is subjected to first pooling processing to obtain a reference image; corresponding weights are given to the pixels of the image to be processed according to the reference image; In this embodiment of the present application, the pixels at different positions of the image to be processed are given weights according to the weight formula, and the image to be processed and the reference image are subjected to a second pooling process according to the weights to obtain the final target image, thereby retaining more image details and making more It conforms to the subjective perception of human vision.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only for the present application. In some embodiments, for those of ordinary skill in the art, other drawings can also be obtained according to these drawings without any creative effort.

1 is a schematic flowchart of an image pooling method provided by an embodiment of the present application;

FIG. 2 is a schematic structural diagram of an image pooling apparatus provided by an embodiment of the present application;

FIG. 3 is a schematic structural diagram of a terminal device provided by an embodiment of the present application.

Detailed ways

In the following description, for the purpose of illustration rather than limitation, specific details such as specific system structures and technologies are set forth in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be practiced in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

It is to be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integers, steps, operations, elements and/or components, but does not exclude one or more other The presence or addition of features, integers, steps, operations, elements, components and/or sets thereof.

It should also be understood that, as used in this specification and the appended claims, the term "and/or" refers to and including any and all possible combinations of one or more of the associated listed items.

As used in the specification of this application and the appended claims, the term "if" may be contextually interpreted as "when" or "once" or "in response to determining" or "in response to detecting ". Similarly, the phrases "if it is determined" or "if the described condition or event is detected" may be interpreted, depending on the context, to mean "once it is determined" or "in response to the determination" or "once the described condition or event is detected" or " In response to detection of the described condition or event".

In addition, in the description of the specification of the present application and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be construed as indicating or implying relative importance.

References in this specification to "one embodiment" or "some embodiments" and the like mean that a particular feature, structure or characteristic described in connection with the embodiment is included in one or more embodiments of the present application. Thus, appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in other embodiments," etc. in various places in this specification are not necessarily All refer to the same embodiment, but mean "one or more but not all embodiments" unless specifically emphasized otherwise. The terms "including", "including", "having" and their conjugations all mean "including but not limited to" unless specifically emphasized otherwise.

The technical solutions provided by the embodiments of the present application are described in detail below.

Referring to a schematic flowchart of an image pooling method shown in FIG. 1, the method can be applied to terminal equipment, and the terminal equipment can be a mobile phone, tablet or computer, etc., and the type of the terminal equipment is not limited here. The method may include the following steps:

Step S101: Acquire the image I to be processed.

Step S102: performing the first pooling process on the to-be-processed image I to obtain a reference image, wherein the to-be-processed image can be a face image, a landscape image, or a car image, etc., and the format of the to-be-processed image can be JPEG, TIFF, RAW, etc., This application does not make any limitation on this. There are two ways to get the reference image:

The first type: First, the image I to be processed is down-sampled to obtain the down-sampled image I _B .

A box filter can be used to downsample the image I to be processed, and the size of the obtained downsampled image is determined by the size ratio between the target image and the image to be processed, such as 1/2 or 1/4 of the length and width of the image to be processed. . It can be understood that the method of nearest neighbor interpolation, bilinear interpolation and trilinear interpolation can also be used to downsample the image I to be processed to obtain a downsampled image. The target image is the final desired image.

In this embodiment of the present application, by performing downsampling, the high-frequency components in the to-be-processed image are weakened to obtain low-frequency components, which is beneficial to obtain global information of the to-be-processed image.

Then, the down-sampled image _IB is smoothed using the filter template to obtain the reference image I~.

The filter template can be an approximate Gaussian filter template, a bilateral filter template, or a guided filter template. When using an approximate Gaussian filter template, you can choose a 3*3 or 5*5 matrix. When exemplarily selecting a 3*3 approximate Gaussian filter template, the generation formula of the reference image is as follows:

The second: obtain a reference image according to the learnable parameters and the image to be processed, and the formula for obtaining the reference image is:

where F(q) represents the learnable parameters of the neural network on the neighborhood Ω _p .

In one embodiment, the image to be processed is input into a neural network to generate a reference image, where the learnable parameter of the neural network is F(q), and the image to be processed maps a neighborhood Ω _p on the original image as a reference through the neural network A pixel on the image to generate a reference image.

Ω _p represents a neighborhood on the image to be processed, the size of the neighborhood is generally 2*2 or 3*3, the neighborhood can contain multiple pixels, and q represents a pixel in a neighborhood Ω _p on the image to be processed point, p represents the pixel point in the reference image corresponding to the neighborhood, that is, the neighborhood corresponds to the pixel point p on the reference image, and F(q) represents the learnable parameter of the neural network on the neighborhood Ω _p .

According to the learnable parameters and the image to be processed, before obtaining the reference image, it also includes training the neural network to obtain the learnable parameters and/or trainable parameters of the neural network, which mainly includes two stages:

Forward propagation stage: The stage in which data is propagated from low-level to high-level, that is, the forward propagation stage.

Backpropagation stage: When the results obtained by forward propagation are inconsistent with expectations, the error is propagated from high-level to low-level for training, that is, the backpropagation stage. The specific training process is:

The first step: the neural network initializes the weights. The parameters of F(q) are all initialized to 0, α and λ are initialized to 1, and then a zero-mean Gaussian perturbation is added to all parameters to break the symmetry.

Step 2: The input data is forwarded through the convolutional layer, the pooling layer, and the fully connected layer to obtain the output value. The input data includes images to be processed, as well as arbitrary images and feature maps.

Step 3: Find the error between the output value of the neural network and the target value, where the target value is set by people according to their needs.

Step 4: When the error is greater than our expected value, the error is transmitted back to the network, and the training is propagated from the high level to the low level, that is, the error of the fully connected layer, the pooling layer, and the convolutional layer is obtained in turn in the backpropagation stage. . The error of each layer can be understood as how much it should bear for the total error of the neural network. When the error is less than or equal to our expected value, the weights are updated according to the error and the training ends.

Step S103: Assign corresponding weights to the pixels of the image to be processed according to the reference image.

与权重成正比，

的差值越大权重越大。其中，q代表待 处理图像上的一个邻域Ω_p内的一个像素点，p代表参考图像中对应于邻域的像 素点，ω_α,λ[p,q]代表像素点q的权重。Different weights ω _{α, λ} [p, q] are assigned to the pixel points q at different positions in the neighborhood Ω _p on the image to be processed, and the greater the difference between the pixel value of the pixel point q and the pixel value of the pixel point p, the greater the The greater the weight corresponding to q, the greater the difference between the pixel value of the pixel point q and the pixel value of the pixel point p, and the greater the weight corresponding to the pixel point q, that is,

proportional to the weight,

The greater the difference, the greater the weight. Among them, q represents a pixel point in a neighborhood Ω _p on the image to be processed, p represents the pixel point corresponding to the neighborhood in the reference image, and ω _α,λ [p, q] represents the weight of the pixel point q.

Exemplarily, the formula for the weight is as follows:

为参考图像中 像素点P的像素值，ε为固定常数，α和λ为神经网络的可训练参数，n为自然 数，ε是一个很小的固定常数，具体数值可根据实际情况确定，α,λ都是可训练 参数，在神经网络训练过程中自适应的调整。通过优化logα和logλ来保证α和λ 都是正数。In this formula, I[q] represents the pixel value of the pixel q in the image to be processed,

is the pixel value of the pixel point P in the reference image, ε is a fixed constant, α and λ are the trainable parameters of the neural network, n is a natural number, ε is a small fixed constant, the specific value can be determined according to the actual situation, α, λ are all trainable parameters, which are adaptively adjusted during the neural network training process. Ensure that both α and λ are positive numbers by optimizing logα and logλ.

Step S104: Perform a second pooling process on the image to be processed and the reference image according to the weight to obtain a target image, where the target image is a pooled image of the image to be processed.

与权重成正比，

的差值越大权重越大。像 素值差值越大，图像细节区域的颜色越丰富，本申请实施例通过对细节区域赋 予较大的权重，故图像细节区域对最后输出的目标图像的影响和贡献更大，从 而保留更多的图像细节，更符合人类视觉的主观感受。In this embodiment of the present application, the final target image is obtained by performing a second pooling process on the to-be-processed image and the reference image according to the weights assigned to the pixels at different positions of the to-be-processed image in step S103. For example, by assigning weights ω _{α, λ} [p, q] to the pixel points q at different positions of the image to be processed according to the pixel values, and according to the weights ω _{α, λ} [p, q], the image to be processed and the reference image are subjected to the second pooling process , to obtain the output O(I)[p] of the final target image, where O(I)[p] represents the pixel value of the pixel corresponding to the P pixel position in the target image. The greater the difference between the pixel value of the pixel point q and the pixel value of the pixel point p, the greater the weight corresponding to the pixel point q, that is

proportional to the weight,

The greater the difference, the greater the weight. The larger the pixel value difference is, the richer the color of the image detail area. By assigning a larger weight to the detail area in this embodiment of the present application, the image detail area has a greater impact and contribution to the final output target image, thereby retaining more The image details are more in line with the subjective perception of human vision.

In one embodiment, the pooling processing formula is further determined according to the weight; according to the pooling processing formula, the second pooling processing is performed on the image to be processed and the reference image to obtain the target image. For example, the reference image and the image to be processed are input into the neural network, and the second pooling process is performed through a detail preserve pooling filter to obtain the target image.

Among them, optional, the pooling processing formula is:

Further, the embodiment of the present application further retains more image details through the second pooling filtering that retains details, which is more in line with the subjective perception of human vision, and is also more conducive to the next image processing, such as image segmentation, image Classification, image difference or image compression, etc.

Although the noise will be over-enhanced, the existence of the trainable parameters α, λ just limits the over-enhancement of the noise.

It should be noted that those skilled in the art are within the technical scope disclosed by the present invention, and other sorting schemes that can be easily thought of should also be within the protection scope of the present invention, and will not be repeated here.

Referring to Fig. 2, it is a schematic diagram of an image pooling apparatus provided by an embodiment of the present application. For the convenience of description, only the part relevant to the embodiment of the present invention is shown, including:

an acquisition module 21, used for acquiring the image to be processed;

The first pooling module 22 is configured to perform a first pooling process on the image to be processed to obtain a reference image;

The weight module 23 is used to assign corresponding weights to the pixels of the image to be processed according to the reference image;

The second pooling module 24 is configured to perform a second pooling process on the image to be processed and the reference image according to the weight to obtain a target image, where the target image is a pooled image of the image to be processed.

The weight module 23 is also used to assign different weights to the pixel points q at different positions in the neighborhood Ω _p on the image to be processed. The greater the weight of , where q represents a pixel in a neighborhood Ω _p on the image to be processed, and p represents a pixel in the reference image corresponding to the neighborhood.

Optionally, the formula for the weight is:

为参考图像中像素点P的像素值，ε为固定常数， α和λ为神经网络的可训练参数，n为自然数。Among them, q represents a pixel point in a neighborhood Ω _p on the image to be processed, p represents the pixel point corresponding to the neighborhood in the reference image, ω _α,λ [p, q] represents the weight of the pixel point q, I [q] represents the pixel value of the pixel point q in the image to be processed,

is the pixel value of the pixel point P in the reference image, ε is a fixed constant, α and λ are the trainable parameters of the neural network, and n is a natural number.

The first pooling module 22 is also used to obtain a reference image according to the learnable parameter and the image to be processed, wherein the formula for obtaining the reference image is:

where F(q) represents the learnable parameters of the neural network on the neighborhood Ω _p .

The second pooling module 24 is further configured to determine the pooling processing formula according to the weight; according to the pooling processing formula, the second pooling processing is performed on the image to be processed and the reference image to obtain the target image;

Among them, optional, the pooling processing formula is:

Among them, O(I)[p] represents the pixel value of the pixel point corresponding to the P pixel point in the target image.

Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example for illustration. In practical applications, the above-mentioned functions can be allocated by different functional units and modules as required. , that is, dividing the internal structure of the mobile terminal into different functional units or modules to complete all or part of the functions described above. Each functional module in the embodiment may be integrated in one processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit, and the above-mentioned integrated units may be implemented in the form of hardware. , and can also be implemented in the form of software functional units. In addition, the specific names of each functional module are only for the convenience of distinguishing from each other, and are not used to limit the protection scope of the present application. For the specific working process of the module in the above-mentioned mobile terminal, reference may be made to the corresponding process in the foregoing method embodiments, which will not be repeated here.

FIG. 3 is a schematic diagram of a terminal device provided by an embodiment of the present application. As shown in FIG. 3 , the terminal device 3 of this embodiment includes a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30. When the processor 30 executes the computer program 32, the steps of the above-mentioned image pooling method are realized, such as steps 101 to 104 shown in FIG. 1 . Alternatively, when the processor 30 executes the computer program 32, the functions of the modules/units in the above-mentioned various apparatus embodiments are realized, for example, the functions of the modules 21 to 24 shown in FIG. 3 .

Exemplarily, the computer program 32 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 31 and executed by the processor 30 to accomplish the present invention. One or more modules/units may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program 32 in the terminal device 3.

The terminal device 3 may be a computing device such as a desktop computer, a notebook, and a palmtop computer. The terminal device may include, but is not limited to, the processor 30 and the memory 31. Those skilled in the art can understand that FIG. 3 is only an example of the terminal device 3 , and does not constitute a limitation on the terminal device 3 , and may include more or less components than shown, or combine some components, or different components For example, the terminal device may also include an input and output device, a network access device, a bus, and the like.

The so-called processor 30 may be a central processing unit (Central Processing Unit, CPU), and may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), Off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 31 may be an internal storage unit of the terminal device 3, such as a hard disk or a memory of the terminal device 3. The memory 31 can also be an external storage device of the terminal device 3, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash memory card (Flash card) equipped on the terminal device 3. Card), etc. Further, the memory 31 may also include both an internal storage unit of the terminal device 3 and an external storage device. The memory 31 is used to store computer programs and other programs and data required by the terminal device. The memory 31 can also be used to temporarily store data that has been output or is to be output.

Embodiments of the present application further provide a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the foregoing method embodiments can be implemented.

The embodiments of the present application provide a computer program product, when the computer program product runs on a mobile terminal, so that the steps in the foregoing method embodiments can be implemented when the mobile terminal executes.

Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. Module completion means dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment may be integrated in one processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit, and the above-mentioned integrated units may adopt hardware. It can also be realized in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing from each other, and are not used to limit the protection scope of the present invention. For the specific working processes of the units and modules in the above system, reference may be made to the corresponding processes in the foregoing method embodiments, and details are not described herein again.

In the above-mentioned embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may implement the described functionality using different methods for each particular application, but such implementations should not be considered beyond the scope of the present invention.

In the embodiments provided by the present invention, it should be understood that the disclosed apparatus/terminal device and method may be implemented in other manners. For example, the apparatus/terminal device embodiments described above are only illustrative. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components. May be combined or may be integrated into another system, or some features may be omitted, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, which may be electrical, mechanical or other forms.

Units described as separate components may or may not be physically separated, and components shown as units may or may not be physical units, that is, may be located in one place, or may be distributed over multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware, and can also be implemented in the form of software functional units.

The integrated module/unit, if implemented as a software functional unit and sold or used as a stand-alone product, may be stored on a computer-readable storage medium. Based on this understanding, the present invention can implement all or part of the processes in the methods of the above embodiments, and can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program is in When executed by the processor, the steps of the foregoing method embodiments can be implemented. The computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying computer program code, recording medium, U disk, removable hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access Memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to the legislation and patent practice, the computer-readable medium does not include It is an electrical carrier signal and a telecommunication signal.

The above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The recorded technical solutions are modified, or some technical features thereof are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the present invention. within the scope of protection.

Claims (14)

1. an image pooling method, is characterized in that, comprises: Get the image to be processed; performing a first pooling process on the to-be-processed image to obtain a reference image; Assign corresponding weights to the pixels of the to-be-processed image according to the reference image; According to the weight, a second pooling process is performed on the to-be-processed image and the reference image to obtain a target image. 2. The method according to claim 1, wherein, according to the reference image, assigning corresponding weights to the pixels of the to-be-processed image, comprising: Different weights are given to the pixel points q at different positions in the neighborhood Ω _p on the image to be processed, and the greater the difference between the pixel value of the pixel point q and the pixel value of the pixel point p, the greater the The larger the weight is, where q represents a pixel point in a neighborhood Ω _p on the image to be processed, and p represents a pixel point in the reference image corresponding to the neighborhood. 3. The method according to claim 1, wherein, according to the reference image, assigning corresponding weights to the pixels of the to-be-processed image, comprising:

Wherein, q represents a pixel in a neighborhood Ω _p on the image to be processed, p represents a pixel in the reference image corresponding to the neighborhood, ω _α,λ [p,q] represents a pixel The weight of point q, I[q] represents the pixel value of the pixel point q in the image to be processed, I～[p] is the pixel value of the pixel point P in the reference image, ε is a fixed constant, α and λ are Trainable parameters of the neural network, n is a natural number. 4. The method according to claim 1, wherein, according to the weight, performing a second pooling process on the to-be-processed image and the reference image to obtain a target image, comprising: Determine the pooling processing formula according to the weight; According to the pooling processing formula, a second pooling process is performed on the to-be-processed image and the reference image to obtain a target image; Wherein, the pooling processing formula is:

Among them, O(I)[p] represents the pixel value of the pixel point corresponding to the P pixel point in the target image, q represents a pixel point in a neighborhood Ω _p on the to-be-processed image, and p represents For the pixels in the reference image corresponding to the neighborhood, ω _α,λ [p,q] represents the weight of the pixel q, and I[q] represents the pixel value of the pixel q in the image to be processed. 5. The method according to claim 1, wherein performing a first pooling process on the image to be processed to obtain a reference image comprises: According to the learnable parameters and the to-be-processed image, a reference image is obtained, wherein the formula for obtaining the reference image is:

in,

is the pixel value of the pixel point P in the reference image, p represents the pixel point in the reference image corresponding to the neighborhood, F(q) represents the learnable parameter of the neural network on the neighborhood Ω _p , q represents A pixel point in a neighborhood Ω _p on the image to be processed, I[q] represents the pixel value of the pixel point q in the image to be processed. 6. The method according to claim 5, wherein, before obtaining the reference image according to the learnable parameter and the to-be-processed image, the method further comprises: The neural network is trained to obtain learnable parameters of the neural network and/or the trainable parameters. 7. The method of claim 1, wherein performing a first pooling process on the to-be-processed image to obtain a reference image comprises: down-sampling the to-be-processed image to obtain a down-sampled image; The downsampled image is smoothed using a filter template to obtain a reference image. 8. An image pooling device, comprising: The acquisition module is used to acquire the image to be processed; a first pooling module, configured to perform a first pooling process on the to-be-processed image to obtain a reference image; a weighting module, configured to assign corresponding weights to the pixels of the to-be-processed image according to the reference image; A second pooling module, configured to perform a second pooling process on the to-be-processed image and the reference image according to the weight to obtain a target image. 9. The apparatus according to claim 8, wherein the weighting module is further configured to assign different weights to pixel points _q at different positions in the neighborhood Ωp on the image to be processed, and the pixel points The greater the difference between the pixel value of q and the pixel value of pixel p, the greater the weight corresponding to pixel q, where q represents a pixel in a neighborhood Ω _p on the image to be processed, and p represents pixels in the reference image corresponding to the neighborhood. 10. The apparatus according to claim 8, wherein the weighting module is further configured to assign corresponding weights to the pixels of the to-be-processed image according to the reference image, wherein:

Wherein, q represents a pixel in a neighborhood Ω _p on the image to be processed, p represents a pixel in the reference image corresponding to the neighborhood, ω _α,λ [p,q] represents a pixel The weight of the point q, I[q] represents the pixel value of the pixel point q in the image to be processed,

is the pixel value of the pixel point P in the reference image, ε is a fixed constant, α and λ are the trainable parameters of the neural network, and n is a natural number. 11. The apparatus of claim 8, wherein the first pooling module is further configured to obtain a reference image according to the learnable parameter and the to-be-processed image, wherein the formula for obtaining the reference image is:

in,

is the pixel value of the pixel point P in the reference image, p represents the pixel point in the reference image corresponding to the neighborhood, F(q) represents the learnable parameter of the neural network on the neighborhood Ω _p , q represents A pixel point in a neighborhood Ω _p on the image to be processed, I[q] represents the pixel value of the pixel point q in the image to be processed. 12. The apparatus according to claim 8, wherein the second pooling module is further configured to determine a pooling processing formula according to the weight; The reference image is subjected to a second pooling process to obtain a target image; Wherein, the pooling processing formula is:

Among them, O(I)[p] represents the pixel value of the pixel point corresponding to the P pixel point in the target image, q represents a pixel point in a neighborhood Ω _p on the to-be-processed image, and p represents For the pixels in the reference image corresponding to the neighborhood, ω _α,λ [p,q] represents the weight of the pixel q, and I[q] represents the pixel value of the pixel q in the image to be processed. 13. A terminal device, characterized in that the terminal device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, when the processor executes the computer program Steps for implementing an image pooling method as claimed in any one of claims 1 to 7. 14. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the image pool according to any one of claims 1 to 7 is realized steps of the method.

Images