Image enhancement method and device, network equipment and storage medium

By using channel attention module and quantum feature fusion network in image enhancement, the problems of loss of detail information and high computational complexity during image enhancement are solved, and efficient recovery and quality improvement of low-illumination images are achieved.

CN118941472APending Publication Date: 2024-11-12CHINA MOBILE (SUZHOU) SOFTWARE TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410986477.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The prior art is prone to problems such as loss of detail information, high computational complexity, and large amount of computing resources during image enhancement. Especially when processing low-illumination images, it is difficult to effectively restore the texture details and brightness information of the image.

Method used

The channel attention module is used to adaptively weighted adjustment of multi-color channel dimensions in the lighting feature extraction network, combined with the reflection feature extraction network and the quantum feature fusion network, and fuse the image light and reflective features through the quantum convolution neural network, and use the Multi-scale fusion strategy and upsampling layer perform image enhancement, and loss function is constructed for supervised training.

Benefits of technology

It improves the accuracy and efficiency of image enhancement, effectively restores the texture details and brightness information of the image, reduces the complexity of the network model, significantly improves the computing efficiency, and quickly restores high-quality images.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941472A_ABST
    Figure CN118941472A_ABST
Patent Text Reader

Abstract

The invention discloses an image enhancement method and device, network equipment and a storage medium, and the method comprises the steps: carrying out the adaptive weighting adjustment of an extracted multi-color channel dimension through a channel attention module when a to-be-enhanced image is subjected to feature extraction through an illumination feature extraction network, and obtaining image illumination features of different scales; performing feature extraction on the to-be-enhanced image through the reflection feature extraction network to obtain image reflection features of different scales; fusing the image illumination features and the image reflection features through a quantum feature fusion network to obtain an enhanced target image; the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network. According to the scheme of the invention, the channel attention module is arranged in the illumination feature extraction network, the advantages of nonlinearity and locality of classical convolution and high parallelism of quantum convolution are combined, and a high-quality image can be quickly and effectively recovered.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of quantum computing technology, and specifically to an image enhancement method, apparatus, network equipment and storage medium. Background Art

[0002] Images acquired under dim light conditions suffer from visual degradation problems such as low visibility, low contrast, and strong noise. Low-light image enhancement technology can improve image quality, thereby improving the accuracy and efficiency of computer vision tasks.

[0003] Traditional image enhancement technology can improve image quality simply and effectively, but it is prone to problems such as loss of image detail information, artifacts, color distortion, and over-enhancement. In addition, the computational complexity of classic convolutional neural networks is high when processing large-scale images and complex convolution kernels, consuming a large amount of computing resources and having low image processing efficiency. Summary of the invention

[0004] At least one embodiment of the present application provides an image enhancement method, apparatus, network device and storage medium, which are used to solve the problem of easy loss of detail information or high computational complexity and consumption of computing resources in the process of image enhancement in the prior art.

[0005] In order to solve the above technical problems, this application is implemented as follows:

[0006] In a first aspect, an embodiment of the present application provides an image enhancement method, comprising:

[0007] When extracting features from the enhanced image through the illumination feature extraction network, a channel attention module is used to adaptively weight the extracted multi-color channel dimensions to obtain image illumination features of different scales; and

[0008] The reflection feature extraction network is used to extract features of the image to be enhanced, and the reflection features of images at different scales are obtained;

[0009] The image illumination feature and the image reflection feature are fused through a quantum feature fusion network to obtain an enhanced target image;

[0010] Among them, the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network.

[0011] Furthermore, the image illumination features include:

[0012] A first local feature and a first global feature, wherein the first local feature is obtained by extracting features using a single-qubit gate through the illumination feature extraction network of the channel attention module, and the first global feature is obtained by extracting features using a double-qubit gate through the illumination feature extraction network of the channel attention module;

[0013] The image reflection characteristics include:

[0014] The second local feature is obtained by extracting features through the reflection feature extraction network using a single-qubit gate, and the second global feature is obtained by extracting features through the reflection feature extraction network using a double-qubit gate.

[0015] Furthermore, the image illumination feature and the image reflection feature are fused through a quantum feature fusion network to obtain an enhanced target image, including:

[0016] A multi-scale fusion strategy is adopted to fuse the image illumination features of different scales and the image reflection features of different scales through interlaced single-bit quantum gates and double-bit quantum gates to obtain the target image.

[0017] Furthermore, the image enhancement method further comprises:

[0018] The target image is enlarged and enhanced by using an upsampling layer and a classical convolutional layer of a quantum convolutional neural network.

[0019] Furthermore, the image enhancement method further comprises:

[0020] The size of the image to be enhanced is adjusted by using the quantum convolutional neural network and a linear interpolation method.

[0021] Furthermore, the image enhancement method further comprises:

[0022] Constructing a loss function of the quantum convolutional neural network;

[0023] Performing supervised training on the quantum convolutional neural network using mean square error according to the loss function;

[0024] Among them, the loss functions include: reflectance similarity loss function, illumination consistency loss function and structure similarity loss function.

[0025] Furthermore, the loss function is:

[0026]

[0027] Among them, α, β, and γ are weight parameters; is the reflectivity similarity loss function; represents the lighting consistency loss function; represents the structural similarity loss function;

[0028]

[0029] N represents the number of network branches for extracting features of the image to be enhanced; R h represents the reflection component corresponding to the image to be enhanced; L h represents the illumination component corresponding to the image to be enhanced; Upsampling represents the upsampling operation; I e I represents the target image; h represents the image to be enhanced.

[0030] In a second aspect, an embodiment of the present application provides an image enhancement device, including:

[0031] An adjustment module, used for adaptively weighting and adjusting the extracted multi-color channel dimensions by using a channel attention module when extracting features of the image to be enhanced through the illumination feature extraction network, so as to obtain image illumination features of different scales; and

[0032] An extraction module is used to extract features of the image to be enhanced through a reflection feature extraction network to obtain image reflection features of different scales;

[0033] A fusion module, used for fusing the image illumination feature and the image reflection feature through a quantum feature fusion network to obtain an enhanced target image;

[0034] Among them, the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network.

[0035] In a third aspect, an embodiment of the present application provides a network device, including a processor, wherein:

[0036] The processor is used to use a channel attention module to adaptively weight and adjust the extracted multi-color channel dimensions when extracting features from the image to be enhanced through the illumination feature extraction network, so as to obtain image illumination features of different scales; and

[0037] The reflection feature extraction network is used to extract features of the image to be enhanced, and the reflection features of images at different scales are obtained;

[0038] The image illumination feature and the image reflection feature are fused through a quantum feature fusion network to obtain an enhanced target image;

[0039] Among them, the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network.

[0040] In a fourth aspect, an embodiment of the present application provides a network device, comprising: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the method described in the first aspect.

[0041] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a program is stored. When the program is executed by a processor, the steps of the method described in the first aspect are implemented.

[0042] According to a sixth aspect, a computer program product is provided, comprising computer instructions, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0043] Compared with the prior art, the image enhancement method, device, network device and storage medium provided in the embodiments of the present application, by setting the channel attention module in the illumination feature extraction network in the quantum convolutional neural network, can adaptively weighted adjust the multi-color channel dimensions of the features extracted from the enhanced image, and can improve the perception ability of the illumination feature extraction network, thereby utilizing the channel relationship between the features to generate a channel attention map to adaptively extract the image illumination features, improve the accuracy of image enhancement, better restore the texture details and brightness information of the image, and combine the advantages of classical convolution nonlinearity, locality and high parallelism of quantum convolution, can effectively extract image features while reducing the complexity of the network model, significantly improving the computational efficiency of the model, thereby quickly and effectively restoring high-quality images. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0045] Figure 1 A schematic diagram of the steps of the image enhancement method according to an embodiment of the present application;

[0046] Figure 2 A schematic diagram of the circuits of the quantum convolution layer and the quantum pooling layer in the quantum convolutional neural network of an embodiment of the present application;

[0047] Figure 3 This is a schematic diagram of the circuit of the quantum feature fusion module in the quantum convolutional neural network of an embodiment of the present application;

[0048] Figure 4 A schematic diagram of the structure of an image enhancement device according to an embodiment of the present application;

[0049] Figure 5 A schematic diagram of the structure of a network device according to an embodiment of the present application;

[0050] Figure 6 This is a schematic diagram of the structure of a network device according to another embodiment of the present application. DETAILED DESCRIPTION

[0051] The exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0052] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable in appropriate circumstances, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, the process, method, system, product or equipment comprising a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment. "And / or" in the specification and claims represents at least one of the connected objects.

[0053] As described in the background technology, in the prior art, image enhancement technology can simply and effectively improve image quality, but it is prone to problems such as loss of image detail information, artifacts, color distortion, over-enhancement, etc. In addition, the computational complexity of the classic convolutional neural network when processing large-scale images and complex convolution kernels is high, the amount of computing resources consumed is large, and the image processing efficiency is low. In order to solve at least one of the above problems, the embodiments of the present application provide an image enhancement method, device, network device and storage medium to improve the quality and efficiency of image enhancement.

[0054] Please refer to Figure 1 , an image enhancement method provided in an embodiment of the present application comprises the following steps:

[0055] Step 101, when extracting features of the image to be enhanced through the illumination feature extraction network, a channel attention module is used to adaptively weight and adjust the extracted multi-color channel dimensions to obtain image illumination features of different scales; and

[0056] Step 102, extracting features of the image to be enhanced through a reflection feature extraction network to obtain image reflection features of different scales;

[0057] Step 103, fusing the image illumination feature and the image reflection feature through a quantum feature fusion network to obtain an enhanced target image;

[0058] Among them, the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network.

[0059] It should be noted that the image to be enhanced is a low-illumination image acquired under dark light conditions and having visual degradation problems such as low visibility, low contrast and strong noise.

[0060] Quantum computers are more conducive to large-scale computing due to their quantum superposition, entanglement characteristics and high parallelism, providing potential exponential acceleration, and can be applied to a variety of computer tasks such as enhanced processing of low-light images.

[0061] It should be noted that a channel attention module (CAM) is introduced into the illumination feature extraction network to utilize the inter-channel relationship of features to generate a channel attention map to adaptively extract image illumination information.

[0062] It should be noted that, based on the Retinex theoretical model, the image I is composed of the illumination image L and the reflection image R, which is expressed as I = R·L. Therefore, in the encoding network in the quantum convolutional neural network, I is input into the illumination feature extraction network and the reflection feature extraction network respectively.

[0063] In one embodiment of the present invention, Figure 2 As shown in the figure, three reflection components R1, R2, R3 and illumination components L1, L2, L3 of different scales are obtained through the three branches E1, E2, and E3 in the network. The input image size of each branch is H×W, H / 2×W / 2, and H / 4×W / 4, and it consists of a quantum convolution layer and a quantum pooling layer.

[0064] In the embodiment of the present invention, when the illumination feature extraction network is used to extract features from the image to be enhanced, a channel attention module is used to adaptively weight the extracted multi-color channel dimensions to obtain image illumination features of different scales, which can be expressed as:

[0065]

[0066] The reflection feature extraction network is used to extract features from the enhanced image to obtain reflection features of images at different scales, which can be expressed as:

[0067]

[0068] Where n = {1, 2, 3} is the network branch number, represents the quantum pooling operation, Represents the quantum convolution operation; CAM represents the channel attention module calculation process.

[0069] The image enhancement method of the embodiment of the present application, by setting the channel attention module in the illumination feature extraction network in the quantum convolutional neural network, can adaptively weighted adjust the multi-color channel dimensions of the features extracted from the enhanced image, can improve the perception ability of the illumination feature extraction network, and thereby utilize the channel relationship between the features to generate a channel attention map to adaptively extract the image illumination features, improve the accuracy of image enhancement, better restore the texture details and brightness information of the image, and combine the advantages of the nonlinearity and locality of classical convolution and the high parallelism of quantum convolution, can effectively extract image features while reducing the complexity of the network model, significantly improve the computational efficiency of the model, and thus quickly and effectively restore high-quality images.

[0070] Optionally, the image illumination feature includes:

[0071] A first local feature and a first global feature, wherein the first local feature is obtained by extracting features using a single-qubit gate through the illumination feature extraction network of the channel attention module, and the first global feature is obtained by extracting features using a double-qubit gate through the illumination feature extraction network of the channel attention module;

[0072] The image reflection characteristics include:

[0073] The second local feature is obtained by extracting features through the reflection feature extraction network using a single-qubit gate, and the second global feature is obtained by extracting features through the reflection feature extraction network using a double-qubit gate.

[0074] Alternatively, if Figure 2 As shown, the quantum circuit is connected by a revolving door R y (θ) and revolving door R z(θ), extracts local and global features of the image through single-qubit gates and double-qubit gates, and uses a pooling layer to perform sampling on the features to reduce redundant information.

[0075] The image enhancement method of the embodiment of the present invention can improve the extraction accuracy of detail features by extracting local features and global features of the image through the single-qubit gate and the double-qubit gate respectively.

[0076] Optionally, fusing the image illumination feature and the image reflection feature through a quantum feature fusion network to obtain an enhanced target image includes:

[0077] A multi-scale fusion strategy is adopted to fuse the image illumination features of different scales and the image reflection features of different scales through interlaced single-bit quantum gates and double-bit quantum gates to obtain the target image.

[0078] In the embodiment of the present invention, Figure 3 As shown, in order to fully integrate the illumination features and reflection features of the image, a multi-scale fusion strategy and quantum feature fusion network are designed;

[0079] The reflection features R1, R2, R3 and the illumination features L1, L2, L3 of different scales are fused through the quantum feature fusion network. Specifically, the quantum feature fusion network consists of interlaced single-bit quantum gates R z (θ) and the two-bit quantum gate R x (θ), parameterized quantum gates are applied to different quantum bits, and the strong entanglement characteristics of quantum are utilized to fully complete the fusion of features.

[0080] The image enhancement method of the embodiment of the present invention can fully integrate the illumination feature and the reflection feature through the interlaced single-bit quantum gate and the double-bit quantum gate.

[0081] Optionally, the image enhancement method further includes:

[0082] The target image is enlarged and enhanced by using an upsampling layer and a classical convolutional layer of a quantum convolutional neural network.

[0083] It should be noted that, in the scheme of the present invention, the features of the image to be enhanced are first extracted through the pooling layer; then, the illumination features and the reflection features are fused through the quantum feature fusion network to obtain the target image; finally, the target image is further amplified through the upsampling layer, and the target image is further enhanced through the classical convolution layer.

[0084] Optionally, the image enhancement method further includes:

[0085] The size of the image to be enhanced is adjusted by using the quantum convolutional neural network and a linear interpolation method.

[0086] It should be noted that quantum computers in the noisy intermediate-scale quantum (NISQ) era are limited by the number of quantum bits and the depth of quantum circuits, and the image size needs to be adjusted before calculation.

[0087] In the embodiment of the present invention, the image is linearly interpolated to a spatial resolution of 64*64 for network training.

[0088] Optionally, the image enhancement method further includes:

[0089] Constructing a loss function of the quantum convolutional neural network;

[0090] Performing supervised training on the quantum convolutional neural network using mean square error according to the loss function;

[0091] Among them, the loss functions include: reflectance similarity loss function, illumination consistency loss function and structure similarity loss function.

[0092] Optionally, the loss function is:

[0093]

[0094] Among them, α, β, and γ are weight parameters; is the reflectivity similarity loss function; represents the lighting consistency loss function; represents the structural similarity loss function;

[0095]

[0096] N represents the number of network branches for extracting features of the image to be enhanced; R h represents the reflection component corresponding to the image to be enhanced; L h represents the illumination component corresponding to the image to be enhanced; Upsampling represents the upsampling operation; I e I represents the target image; h represents the image to be enhanced.

[0097] The image enhancement method of the embodiment of the present invention, by performing supervised training on the quantum convolutional neural network through the loss function, can further improve the accuracy of the quantum convolutional neural network for the image to be processed.

[0098] like Figure 4As shown, the embodiment of the present application provides an image enhancement device 400, including:

[0099] An adjustment module 401 is used to use a channel attention module to perform adaptive weighted adjustment on the extracted multi-color channel dimensions when extracting features of the image to be enhanced through the illumination feature extraction network, so as to obtain image illumination features of different scales; and

[0100] An extraction module 402 is used to extract features of the image to be enhanced through a reflection feature extraction network to obtain image reflection features of different scales;

[0101] A fusion module 403 is used to fuse the image illumination feature and the image reflection feature through a quantum feature fusion network to obtain an enhanced target image;

[0102] Among them, the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network.

[0103] The image enhancement device of the embodiment of the present application, by setting the channel attention module in the illumination feature extraction network in the quantum convolutional neural network, can adaptively weighted adjust the multi-color channel dimensions of the features extracted from the enhanced image, and can improve the perception ability of the illumination feature extraction network, so as to utilize the channel relationship between the features to generate a channel attention map to adaptively extract the image illumination features, improve the accuracy of image enhancement, better restore the texture details and brightness information of the image, and combine the advantages of the nonlinearity and locality of classical convolution and the high parallelism of quantum convolution, so as to effectively extract image features while reducing the complexity of the network model, significantly improving the computational efficiency of the model, thereby quickly and effectively restoring high-quality images.

[0104] Please refer to Figure 5 , the embodiment of the present application further provides a network device 500, including: a processor 501;

[0105] The processor 501 is used to use a channel attention module to adaptively weight and adjust the extracted multi-color channel dimensions when extracting features of the image to be enhanced through the illumination feature extraction network to obtain image illumination features of different scales; and

[0106] The reflection feature extraction network is used to extract features of the image to be enhanced, and the reflection features of images at different scales are obtained;

[0107] The image illumination feature and the image reflection feature are fused through a quantum feature fusion network to obtain an enhanced target image;

[0108] Among them, the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network.

[0109] Please refer to Figure 6 The embodiment of the present application also provides a network device 600, including a processor 601, a memory 602, and a computer program stored in the memory 602 and executable on the processor 601. When the computer program is executed by the processor 601, each process of the above-mentioned image enhancement method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0110] The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, each process of the above-mentioned image enhancement method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0111] The present application also provides a computer program product, including computer instructions, which, when executed by a processor, implement the above Figure 1 The various processes of the method embodiment shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0112] It should be noted that, in this article, the terms "include", "include" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0113] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0114] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims

1. An image enhancement method, characterized in that: include: When extracting features from the enhanced image through the illumination feature extraction network, a channel attention module is used to adaptively weight the extracted multi-color channel dimensions to obtain image illumination features of different scales; and The reflection feature extraction network is used to extract features of the image to be enhanced, and the reflection features of images at different scales are obtained; The image illumination feature and the image reflection feature are fused through a quantum feature fusion network to obtain an enhanced target image; Among them, the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network.

2. The image enhancement method according to claim 1, characterized in that: The image illumination features include: A first local feature and a first global feature, wherein the first local feature is obtained by extracting features using a single-qubit gate through the illumination feature extraction network of the channel attention module, and the first global feature is obtained by extracting features using a double-qubit gate through the illumination feature extraction network of the channel attention module; The image reflection characteristics include: The second local feature is obtained by extracting features through the reflection feature extraction network using a single-qubit gate, and the second global feature is obtained by extracting features through the reflection feature extraction network using a double-qubit gate.

3. The image enhancement method according to claim 1, characterized in that: The step of fusing the image illumination feature and the image reflection feature through a quantum feature fusion network to obtain an enhanced target image includes: A multi-scale fusion strategy is adopted to fuse the image illumination features of different scales and the image reflection features of different scales through interlaced single-bit quantum gates and double-bit quantum gates to obtain the target image.

4. The image enhancement method according to claim 1, characterized in that: The method further comprises: The target image is enlarged and enhanced by using an upsampling layer and a classical convolutional layer of a quantum convolutional neural network.

5. The image enhancement method according to claim 1, characterized in that: The method further comprises: The size of the image to be enhanced is adjusted by using the quantum convolutional neural network and a linear interpolation method.

6. The image enhancement method according to claim 1, characterized in that: The method further comprises: Constructing a loss function of the quantum convolutional neural network; Performing supervised training on the quantum convolutional neural network using mean square error according to the loss function; Among them, the loss functions include: reflectance similarity loss function, illumination consistency loss function and structure similarity loss function.

7. The image enhancement method according to claim 6, characterized in that: The loss function is: Among them, α, β, and γ are weight parameters; is the reflectance similarity loss function; represents the lighting consistency loss function; represents the structural similarity loss function; N represents the number of network branches for extracting features of the image to be enhanced; R h represents the reflection component corresponding to the image to be enhanced; L h represents the illumination component corresponding to the image to be enhanced; Upsampling represents the upsampling operation; I e I represents the target image; h represents the image to be enhanced.

8. An image enhancement device, characterized in that: include: An adjustment module is used to use a channel attention module to adaptively weight and adjust the extracted multi-color channel dimensions when extracting features of the enhanced image through the illumination feature extraction network, so as to obtain image illumination features of different scales; as well as, An extraction module is used to extract features of the image to be enhanced through a reflection feature extraction network to obtain image reflection features of different scales; A fusion module, used for fusing the image illumination feature and the image reflection feature through a quantum feature fusion network to obtain an enhanced target image; Among them, the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network.

9. A network device, characterized in that: comprising a processor, wherein The processor is used to use a channel attention module to adaptively weight and adjust the extracted multi-color channel dimensions when extracting features from the image to be enhanced through the illumination feature extraction network, so as to obtain image illumination features of different scales; and The reflection feature extraction network is used to extract features of the image to be enhanced, and the reflection features of images at different scales are obtained; The image illumination feature and the image reflection feature are fused through a quantum feature fusion network to obtain an enhanced target image; Among them, the illumination feature extraction network, the reflection feature extraction network and the quantum feature fusion module are network layers in the same quantum convolutional neural network.

10. A network device, characterized in that: include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program implements the steps of the method according to any one of claims 1 to 7 when executed by the processor.

11. 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 steps of the method according to any one of claims 1 to 7 are implemented.

12. A computer program product, characterized in that The method comprises computer instructions, which, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.