Image processing system based on artificial intelligence
By integrating deep learning algorithms and complex mathematical formulas in the image processing system, the problem of traditional image processing technology not ideal for complex image data processing is solved, and efficient and accurate image processing effects are achieved.
Patent Information
- Application Number
- CN202411344338.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-25
- Publication Date
- 2025-03-21
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional image processing technology relies on fixed algorithms and models, and the results of complex and changeable image data processing are not ideal.
Design an image processing system based on artificial intelligence, and realize efficient and accurate processing of images by integrating deep learning algorithms and complex mathematical formulas, including image acquisition modules, preprocessing modules, feature extraction modules, deep learning modules, post-processing modules and output modules.
Through the combination of deep learning module and feature extraction module, efficient processing and accurate analysis of images are achieved, and the working efficiency of the system is improved.
Smart Images
Figure CN119672502A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an image processing system based on artificial intelligence. Background Art
[0002] Artificial intelligence is a theory, method, technology and application system that simulates, extends and expands human intelligence. It belongs to the new technical science. It is an important part of the intelligent discipline, aiming to understand the essence of intelligence and produce an intelligent machine that can respond in a similar way to human intelligence. The research field of artificial intelligence is very broad, including robots, language recognition, image recognition, natural language processing, expert systems, machine learning, computer vision, etc. Its core principle is to use algorithms such as machine learning, deep learning, neural networks, etc. With the rapid development of artificial intelligence technology, image processing has become one of its important application fields.
[0003] However, traditional image processing technology often relies on fixed algorithms and models, and its processing effect is not ideal for complex and changeable image data. Therefore, the present invention proposes an image processing system based on artificial intelligence, which integrates advanced deep learning algorithms and complex mathematical formulas to achieve efficient and accurate processing of images. Summary of the invention
[0004] To solve the above problems, the present invention provides the following technical solutions:
[0005] An image processing system based on artificial intelligence includes an image acquisition module, a preprocessing module, a feature extraction module, a deep learning module, a post-processing module and an output module. The image acquisition module is responsible for receiving image data to be processed and passing it to the preprocessing module; the preprocessing module preprocesses the input image, including operations such as denoising, enhancement and normalization; the feature extraction module further performs feature analysis and processing on the image based on the features extracted by the deep learning module; the deep learning module uses a deep neural network to repair, classify and process the feature-extracted image; the output module outputs the output result of the post-processing module in an appropriate format for user use or further analysis.
[0006] Furthermore, the image acquisition module includes a scanner and a format converter, which is responsible for acquiring the image to be processed represented as X=[X1, X2, ..., Xn], where n is a positive integer, and converting it into a format recognizable by the system. The acquired image data will then be transmitted to other parts of the system for further processing and analysis.
[0007] Furthermore, the preprocessing module includes a processor, which preprocesses the collected image data, including denoising, enhancement, normalization and other operations. The specific steps are as follows: The specific steps of denoising are as follows:
[0008] Denoising: A denoising algorithm based on non-local mean filtering is used. This algorithm denoises by comparing the similarities between pixels. The data after denoising is
[0009] NL[p]=∑q∈S(p)·w(p,q)·I[q]
[0010] Where NL[p] represents the value of pixel p after non-local mean filtering, S(p) is the neighborhood set of pixel p, usually a window or region around p, w(p,q) is the similarity weight between pixels p and q, usually calculated based on the similarity of their grayscale values, and I[q] is the original grayscale value of pixel q.
[0011] Furthermore, the specific steps of the image enhancement processing are as follows:
[0012] Enhancement processing: The histogram equalization algorithm is used for enhancement processing. The contrast of the image is enhanced by stretching the pixel intensity distribution. For pixels with gray level (r), the gray level after mapping is
[0013]
[0014] Where s is the output gray level after mapping; r is the input gray level; L is the total number of gray levels p r (w) is the cumulative distribution function of the gray level r; ξ, ψ, ζ are certain control parameters used to adjust the shape and characteristics of the mapping function.
[0015] Furthermore, the preprocessing module includes image data normalization, that is, normalizing the enhanced image data to a range of [0, 1] to facilitate subsequent processing.
[0016] Furthermore, the specific analysis steps of the feature extraction module are as follows:
[0017] The deep learning algorithm is used to extract features from the preprocessed image. The specific steps are as follows: including convolution operation and pooling operation: The specific analysis steps of the convolution operation are as follows:
[0018] Convolution operation: Convolutional neural network is used for feature extraction. For the input image X, the output (Y) after the convolution layer can be expressed as:
[0019]
[0020] in represents the value at position (i, j) in the output feature map m, (i, j) represents the position of the feature map, X (i+a-1)(j+b-1)c Represents the pixel value of the C channel at position (i+a-1)(j+b-1) in X, W abcm is the weight in the convolution kernel W, a and b are the spatial position index of the convolution kernel, c is the input channel index, and b is the output channel index. m is the bias term, corresponding to the bias of the output feature map m, is the additional noise or error term that may be present due to the uncertainty in modeling the real world.
[0021] Furthermore, the feature extraction module also includes a pooling operation, and the specific analysis steps are as follows:
[0022] Pooling operation: Use maximum pooling or average pooling to reduce the dimension of the feature map after convolution. For the feature map (F), its output (F') after the pooling layer can be expressed as
[0023]
[0024] Where F'(i, j) represents the feature map output, expressing a nonlinear response, R ij It is the pooling window at position (i, j) on the feature map F. It can be a window of any shape, where α, β, γ, and δ are parameters that control the pooling process. α may control the output magnification, β and γ may adjust the weighted influence of the part within the window, and δ may be a power exponent that affects the nonlinear response of the output.
[0025] Furthermore, the deep learning module includes a fully connected layer and an activation function. The features extracted by the feature extraction module are input into the deep learning model for training and learning. The fully connected layer and the activation function are combined to classify, enhance or repair the extracted features. The formula for the combination of the fully connected layer and the activation function is as follows:
[0026]
[0027] Among them, H = [H1, H2, ..., Hn] is the feature vector or feature matrix obtained by standardizing and characterizing it from the feature extraction module X, M is the weight matrix of the fully connected layer, v is the bias vector, is the activation function; the fully connected layer maps feature X to a higher-dimensional feature space, and this formula is used to achieve image classification, enhancement, or restoration.
[0028] Furthermore, the result output module outputs the output results of the deep learning module in the form of images or text for users to view and use.
[0029] Furthermore, an electronic device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements an artificial intelligence-based image processing system as described in any one of claims 1 to 9 when executing the computer program.
[0030] Furthermore, a computer-readable storage medium is characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the image processing system based on artificial intelligence according to any one of claims 1 to 10 is implemented, comprising the following steps:
[0031] Step 1: Using the image acquisition module, the module is responsible for receiving the image data to be processed and passing it to the preprocessing module.
[0032] Step 2: The preprocessing module preprocesses the image transmitted by the image acquisition module, including operations such as denoising, enhancement and normalization;
[0033] Step 3: The feature extraction module further extracts, analyzes and processes the image after being processed by the preprocessing module based on the features extracted by the deep learning module;
[0034] Step 4: The deep learning module uses a deep neural network to repair, classify and process the feature-extracted image;
[0035] Step 5: The output module outputs the output results of the post-processing module in an appropriate format for user use or further analysis.
[0036] Compared with the prior art, the present invention has the following beneficial effects:
[0037] In the present invention, first, the image acquisition module is responsible for receiving the image data to be processed and passing it to the preprocessing module; the preprocessing module preprocesses the input image, including operations such as denoising, enhancement and normalization; the feature extraction module further performs feature analysis and processing on the image based on the features extracted by the deep learning module; the deep learning module uses a deep neural network to repair, classify and process the feature-extracted image; the output module outputs the output results of the post-processing module in an appropriate format for user use or further analysis. The deep learning module and feature extraction module of the present invention use complex mathematical formulas and algorithms to achieve efficient processing and accurate analysis of images, further improving the system work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;
[0039] Figure 1 The figure is a schematic diagram of the steps of an image processing system based on artificial intelligence of the present invention. DETAILED DESCRIPTION
[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0041] like Figure 1 As shown, it includes an image acquisition module, a preprocessing module, a feature extraction module, a deep learning module, a post-processing module and an output module. The image acquisition module is responsible for receiving the image data to be processed and passing it to the preprocessing module; the preprocessing module preprocesses the input image, including operations such as denoising, enhancement and normalization; the feature extraction module further performs feature analysis and processing on the image based on the features extracted by the deep learning module; the deep learning module uses a deep neural network to repair, classify and process the feature-extracted image; the output module outputs the output results of the post-processing module in an appropriate format for user use or further analysis.
[0042] The image acquisition module includes a scanner and a format converter, which is responsible for acquiring the image to be processed represented as X = [X1, X2, ..., Xn], where n is a positive integer, and converting it into a format recognizable by the system. The acquired image data will then be transmitted to other parts of the system for further processing and analysis.
[0043] The preprocessing module includes a processor, which preprocesses the collected image data, including denoising, enhancement and normalization. The specific steps are as follows: The specific steps of denoising are as follows:
[0044] Denoising: A denoising algorithm based on non-local mean filtering is used. This algorithm denoises by comparing the similarities between pixels. The data after denoising is
[0045] NL[p]=∑q∈S(p)·w(p,q)·I[q]
[0046] Where NL[p] represents the value of pixel p after non-local mean filtering, S(p) is the neighborhood set of pixel p, usually a window or region around p, w(p,q) is the similarity weight between pixels p and q, usually calculated based on the similarity of their grayscale values, and I[q] is the original grayscale value of pixel q.
[0047] In the preprocessing module, the specific steps of image enhancement processing are as follows:
[0048] Enhancement processing: The histogram equalization algorithm is used for enhancement processing. The contrast of the image is enhanced by stretching the pixel intensity distribution. For pixels with gray level (r), the gray level after mapping is
[0049]
[0050] Where s is the output gray level after mapping; r is the input gray level; L is the total number of gray levels p r (w) is the cumulative distribution function of the gray level r; ξ, ψ, ζ are certain control parameters used to adjust the shape and characteristics of the mapping function.
[0051] The preprocessing module includes image data normalization, that is, normalizing the enhanced image data to the range of [0, 1] to facilitate subsequent processing.
[0052] The data is processed by the processor of the electronic device to achieve noise reduction, enhancement and normalization of the image, which is beneficial to the later processing of the image.
[0053] The specific analysis steps of the feature extraction module are as follows:
[0054] The deep learning algorithm is used to extract features from the preprocessed image. The specific steps are as follows: including convolution operation and pooling operation: The specific analysis steps of convolution operation are as follows:
[0055] Convolution operation: Convolutional neural network is used for feature extraction. For the input image X, the output (Y) after the convolution layer can be expressed as:
[0056]
[0057] in represents the value at position (i, j) in the output feature map m, (i, j) represents the position of the feature map, X (i+a-1)(j+b-1)c Represents the pixel value of the C channel at position (i+a-1)(j+b-1) in X, W abcm is the weight in the convolution kernel W, a and b are the spatial position index of the convolution kernel, c is the input channel index, and b is the output channel index. m is the bias term, corresponding to the bias of the output feature map m, is the additional noise or error term that may be present due to the uncertainty in modeling the real world.
[0058] The feature extraction module also includes pooling operations. The specific analysis steps are as follows:
[0059] Pooling operation: Use maximum pooling or average pooling to reduce the dimension of the feature map after convolution. For the feature map (F), its output (F') after the pooling layer can be expressed as
[0060]
[0061] Where F'(i, j) represents the feature map output, expressing a nonlinear response, R ij It is the pooling window at position (i, j) on the feature map F. It can be a window of any shape, where α, β, γ, and δ are parameters that control the pooling process. α may control the output magnification, β and γ may adjust the weighted influence of the part within the window, and δ may be a power exponent that affects the nonlinear response of the output.
[0062] The deep learning module includes a fully connected layer and an activation function. The features extracted by the feature extraction module are input into the deep learning model for training and learning. The fully connected layer and the activation function are combined to classify, enhance or repair the extracted features. The formula for the combination of the fully connected layer and the activation function is as follows:
[0063]
[0064] Among them, H = [H1, H2, ..., Hn] is the feature vector or feature matrix obtained by standardizing and characterizing it from the feature extraction module X, M is the weight matrix of the fully connected layer, v is the bias vector, is the activation function; the fully connected layer maps feature X to a higher-dimensional feature space, and this formula is used to achieve image classification, enhancement, or restoration.
[0065] The result output module outputs the output results of the deep learning module in the form of images or text for users to view and use.
[0066] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements an artificial intelligence-based image processing system as claimed in any one of claims 1 to 9 when executing the computer program.
[0067] An image processing system based on artificial intelligence, characterized in that it includes the following steps:
[0068] Step 1: Using the image acquisition module, the module is responsible for receiving the image data to be processed and passing it to the preprocessing module.
[0069] Step 2: The preprocessing module preprocesses the image transmitted by the image acquisition module, including operations such as denoising, enhancement and normalization;
[0070] Step 3: The feature extraction module further extracts, analyzes and processes the image after being processed by the preprocessing module based on the features extracted by the deep learning module;
[0071] Step 4: Deep learning module, using deep neural network to repair, classify and process the feature extracted images;
[0072] Step 5: Output module, which outputs the output results of the post-processing module in an appropriate format for user use or further analysis.
[0073] The working principle of the present invention is as follows: electronic devices are used to control the work of each module, firstly an image acquisition module is used, the image acquisition module is responsible for receiving the image data to be processed, responsible for acquiring the image to be processed represented as X=[X1, X2,..., Xn], n is a positive integer, and passing it to the preprocessing module, the preprocessing module preprocesses the image transmitted by the image acquisition module, including denoising, enhancement, normalization, etc., then the feature extraction module, based on the features extracted by the deep learning module, further extracts features, analyzes and processes the image after being processed by the preprocessing module; the data processed by the feature extraction module is transmitted to the deep learning module, the deep learning module uses a deep neural network to repair, classify and process the feature extracted image, and then the data is output through the output module, the module outputs the output result of the post-processing module in an appropriate format for user use or further analysis.
[0074] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.
Claims
1. An image processing system based on artificial intelligence, characterized in that: It includes an image acquisition module, a preprocessing module, a feature extraction module, a deep learning module, a post-processing module and an output module. The image acquisition module is responsible for receiving the image data to be processed and passing it to the preprocessing module; The preprocessing module performs preprocessing on the input image, including operations such as denoising, enhancement and normalization; The feature extraction module further performs feature analysis and processing on the image based on the features extracted by the deep learning module; the deep learning module uses a deep neural network to repair, classify and process the feature-extracted image; the output module outputs the output results of the post-processing module in an appropriate format for user use or further analysis.
2. The image processing system based on artificial intelligence according to claim 1, characterized in that: The image acquisition module includes a scanner and a format converter, which is responsible for acquiring the image to be processed represented as X=[X1, X2, ..., Xn], where n is a positive integer, and converting it into a format recognizable by the system. The acquired image data will then be transmitted to other parts of the system for further processing and analysis.
3. The image processing system based on artificial intelligence according to claim 2, characterized in that: The preprocessing module includes a processor, which preprocesses the collected image data, including denoising, enhancement, normalization and other operations. The specific steps are as follows: The specific steps of denoising are as follows: Denoising: A denoising algorithm based on non-local mean filtering is used. This algorithm denoises by comparing the similarities between pixels. The data after denoising is NL[p]=∑q∈S(p)·w(p,q)·I[q] Where NL[p] represents the value of pixel p after non-local mean filtering, S(p) is the neighborhood set of pixel p, usually a window or region around p, w(p,q) is the similarity weight between pixels p and q, usually calculated based on the similarity of their grayscale values, and I[q] is the original grayscale value of pixel q.
4. The image processing system based on artificial intelligence according to claim 2, characterized in that: In the preprocessing module, the specific steps of the image enhancement processing are as follows: Enhancement processing: The histogram equalization algorithm is used for enhancement processing. The contrast of the image is enhanced by stretching the pixel intensity distribution. For pixels with gray level (r), the gray level after mapping is Where s is the output gray level after mapping; r is the input gray level; L is the total number of gray levels p r (w) is the cumulative distribution function of the gray level r; ξ, ψ, ζ are certain control parameters used to adjust the shape and characteristics of the mapping function.
5. The image processing system based on artificial intelligence according to claim 4, characterized in that: The preprocessing module includes image data normalization, that is, normalizing the enhanced image data to the range of [0, 1] to facilitate subsequent processing.
6. The image processing system based on artificial intelligence according to claim 3, characterized in that: The specific analysis steps of the feature extraction module are as follows: The deep learning algorithm is used to extract features from the preprocessed image. The specific steps are as follows: including convolution operation and pooling operation: The specific analysis steps of the convolution operation are as follows: Convolution operation: Convolutional neural network is used for feature extraction. For the input image X, the output (Y) after the convolution layer can be expressed as: in represents the value at position (i, j) in the output feature map m, (i, j) represents the position of the feature map, X (i+a-1)(j+b-1)c Represents the pixel value of the C channel at position (i+a-1)(j+b-1) in X, W abcm is the weight in the convolution kernel W, a and b are the spatial position index of the convolution kernel, c is the input channel index, and b is the output channel index. m is the bias term, corresponding to the bias of the output feature map m, is the additional noise or error term that may be present due to the uncertainty in modeling the real world.
7. The image processing system based on artificial intelligence according to claim 6, characterized in that: The feature extraction module also includes a pooling operation, and the specific analysis steps are as follows: Pooling operation: Use maximum pooling or average pooling to reduce the dimension of the feature map after convolution. For the feature map (F), its output (F') after the pooling layer can be expressed as Where F'(i, j) represents the feature map output, expressing a nonlinear response, R ij It is the pooling window at position (i, j) on the feature map F. It can be a window of any shape, where α, β, γ, and δ are parameters that control the pooling process. α may control the output magnification, β and γ may adjust the weighted influence of the part within the window, and δ may be a power exponent that affects the nonlinear response of the output.
8. The image processing system based on artificial intelligence according to claim 6, characterized in that: The deep learning module includes a fully connected layer and an activation function. The features extracted by the feature extraction module are input into the deep learning model for training and learning. The fully connected layer and the activation function are combined to classify, enhance or repair the extracted features. The formula for the combination of the fully connected layer and the activation function is as follows: Among them, H = [H1, H2, ..., Hn] is the feature vector or feature matrix obtained by standardizing and characterizing it from the feature extraction module X, M is the weight matrix of the fully connected layer, v is the bias vector, is the activation function; the fully connected layer maps feature X to a higher-dimensional feature space, and this formula is used to achieve image classification, enhancement, or restoration.
9. The artificial intelligence-based image processing system according to claim 8, characterized in that: The result output module outputs the output results of the deep learning module in the form of images or text for users to view and use.
10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, an artificial intelligence-based image processing system as described in any one of claims 1 to 9 is implemented.
11. An image processing method based on artificial intelligence, using an image processing system based on artificial intelligence as described in any one of claims 1 to 10, characterized in that: The following steps are involved: Step 1: Using the image acquisition module, the module is responsible for receiving the image data to be processed and passing it to the preprocessing module. Step 2: The preprocessing module preprocesses the image transmitted by the image acquisition module, including operations such as denoising, enhancement and normalization; Step 3: The feature extraction module further extracts, analyzes and processes the image after being processed by the preprocessing module based on the features extracted by the deep learning module; Step 4: The deep learning module uses a deep neural network to repair, classify and process the feature-extracted image; Step 5: The output module outputs the output results of the post-processing module in an appropriate format for user use or further analysis.