CN119180818A

CN119180818A - Image complexity calculation method and system considering multi-scale fusion

Info

Publication number: CN119180818A
Application number: CN202411687098.2A
Authority: CN
Inventors: 包兴先; 李江昊; 金晰然; 王茂洁; 郝颖奎; 吴微; 杨硕; 孙昊楠; 马祖亮
Original assignee: China University of Petroleum East China
Current assignee: China University of Petroleum East China
Priority date: 2024-11-25
Filing date: 2024-11-25
Publication date: 2024-12-24
Anticipated expiration: 2044-11-25
Also published as: CN119180818B

Abstract

The invention belongs to the technical field of image processing, and discloses a method and a system for calculating image complexity by considering multi-scale fusion. The method comprises determining whether the pixel values of the input image are the same, and determining the widths of two super-parametric filtersAnd the number of iterationsCalculating the number of iterationsCoarse granularity processing to obtain the first imageReformed image obtained after sub-coarse granularityAnd to the firstReformed image obtained after sub-coarse granularityExpanding to obtain the firstReformed image obtained by sub-coarseningExtended Image. The invention expands the application range of the technology, and makes the technology more suitable for the analysis requirement of the common user on the complexity of the conventional image.

Description

Image complexity calculation method and system considering multi-scale fusion

Technical Field

The invention belongs to the technical field of image processing, and particularly relates to an image complexity calculation method and system considering multi-scale fusion.

Background

Measurement of image complexity has important applications in a number of fields, including computer vision, pattern recognition, image compression, biomedical image analysis, and the like. The quantization of complexity can help not only improve the accuracy of image processing and analysis, but also provide important information about the structure and content of the image. The definition of complexity is generally related to the structural features and content density of an image, reflecting the amount of information contained in the image and its organization.

The prior art methods of computing image complexity are diverse, and involve feature extraction from simple edge detection to deep learning. By the method, the complexity of the image can be quantitatively analyzed, and the processing strategy is optimized and the information processing efficiency is improved according to specific requirements. The image complexity calculation has very key practical application value for improving the image processing technology, enhancing the user experience and optimizing the resource use.

Wherein prior art Bagrov A A, iakovlev I A, iliasov A A, et al . Multiscale structural complexity of natural patterns[J]. Proceedings of the National Academy of Sciences, 2020, 117(48): 30241-30251, provide a complexity calculation method through coarsening. This method first gradually downsamples the original image to produce a series of progressively lower resolution images. The width and height of these images are adjusted to be uniform, and then the complexity of the images is calculated using the initially and finally generated images. By analyzing the variations between images of different resolutions, the method attempts to capture the overall complexity characteristics of the image. Although this method is simple and easy to implement, it has certain limitations.

Through the above analysis, the problems and defects existing in the prior art are as follows:

(1) The method requires the images to be consistent in width and height before processing, and is not suitable for the requirement of a common user on analyzing the complexity of the conventional images;

(2) According to the method, the complexity is calculated by only utilizing the initial and final generated reformation images, and the image information of each scale in the coarse granularity process is ignored, so that the description of the image complexity is not comprehensive enough;

(3) The final complexity calculation result is usually between 0 and 1, and the smaller numerical range makes the complexity difference between different images indistinguishable, so that the effectiveness of the method in practical application is limited;

(4) The method is computationally inefficient in processing high resolution images or large-scale data sets, which becomes a bottleneck in practical applications.

These drawbacks indicate that existing coarsening methods have room for improvement in accuracy, comprehensiveness and computational efficiency.

Disclosure of Invention

In order to overcome the problems in the related art, the embodiment of the invention provides a method and a system for calculating image complexity by considering multi-scale fusion.

The technical scheme is that the image complexity calculating method considering multi-scale fusion comprises the following steps:

S1, judging whether pixel values of the width and the height of an input image are the same, if not, filling an image background, and if the larger value is an odd number, changing the image size +1 into an even number;

S2, determining widths of two super-parametric filters And the number of iterationsCalculating the number of iterationsIs a value range of (a);

S3, performing coarse granularity treatment on the image to sequentially obtain the first Reformed image after sub-coarse granularity;

S4, the second pairReformed image obtained after sub-coarse granularityExpanding to obtain an expanded image;

S5, calculating the firstReformed image obtained after sub-coarse granularityIs a spread image of (a)And the firstReformed image obtained by sub-coarseningDegree of overlap between;

S6, calculating image complexity。

In step S1, after filling the image background, normalizing the image data, wherein the background filling mode is divided into edge filling and positioning filling, the positioning filling is that the original image is placed on the left side of the background filling area for contrast analysis, and the normalization processing calculation formula is as follows:

;

In the formula, Respectively the first image after pretreatmentLine 1The RGB channel normalized values for the column pixels,Respectively the first in the original imageLine 1RGB channel values for column pixels.

In step S2, the filter widthFilter widthAnd the number of iterationsIs two super parameters in the coarse granularity process, and the width and height of the reformed image obtained by coarse granularity are reduced each timeMultiple times, and minimum value of width and height is 1, iteration timesThe value range calculation formula of (2) is:

;

In the formula, Is the firstThe side length of the image obtained by the sub-coarse granularity is an integer,Is the side length of the original image.

In step S3, coarse granularity processing is performed on the image, including:

In the course of coarsening, an image is divided into a plurality of sizes Adding the values of each pixel of the pixel block to obtain an average value to obtain a pixel point, replacing the original pixel block with the pixel point, traversing each pixel block to complete one-time coarse granularity, and the firstReformed image obtained after sub-coarse granularityThe calculation formula is as follows:

;

In the formula, Is the firstSecond coarsely grained reformed imageLine 1Column pixel points; is the first image before coarsening Line 1Column pixel block of the first columnLine 1Column pixel points; Is the first Second coarsely grained reformed imageLine 1And (5) column pixel points.

In step S4, for the firstReformed image obtained after sub-coarse granularityThe nearest neighbor up-sampling expansion is adopted for expansion to obtain the firstReformed image obtained by sub-coarseningIs a spread image of (a)The calculation formula of the nearest neighbor upsampling extension is:

;

In the formula, Is thatFirst, theLine 1The pixel points of the column are arranged,Is thatFirst, theLine 1The pixel points of the column are arranged,Is thatFirst, theLine 1The pixel points of the column are arranged,Is thatFirst, theLine 1And (5) column pixel points.

Further, the image is reformedIs of the side length ofReformed image obtained after sub-coarse granularityA kind of electronic deviceDouble expanding imageAnd (3) withUniform size and satisfiesWherein, the method comprises the steps of,Is the firstThe side length of the image obtained by the sub-coarse granularity is an integer,In order to expand the side length of the image after expansion,Is thatIs used for expanding the image.

In step S5, the first is calculatedReformed image obtained after sub-coarse granularityIs a spread image of (a)And the firstReformed image obtained by sub-coarseningDegree of overlap betweenComprising:

Is that Is the first of (2)Line 1The column of pixels, which is a vector,And (3) withThe product of (2) is a vector product, and the calculation formula is as follows:

;

In the formula, Is the first-1 St coarsening of the obtained reformed imageLine 1And (5) column pixel points.

In step S6, image complexityThe calculation formula of (2) is as follows:

;

In the formula, As a component of the complexity C,Is the firstReformed image obtained by sub-coarseningIs a spread image of (a)And the firstReformed image obtained by sub-coarseningThe overlap ratio between the two.

Further, image complexityObtained by exponential function amplificationThe expression is:

;

In the formula, For the amplified image complexity, C is the original image complexity,Is the base of natural logarithms.

It is another object of the present invention to provide an image complexity calculation system taking into account multi-scale fusion, the system implementing the image complexity calculation method taking into account multi-scale fusion, the system comprising:

The image background filling module is used for judging whether the pixel values of the width and the height of the input image are the same, if not, filling the image background, and if the larger value is an odd number, the image size is +1 and becomes an even number;

The iteration number value range determining module is used for determining the widths of two super-parameter filters And the number of iterationsCalculating the number of iterationsIs a value range of (a);

a reorganization image acquisition module for determining two super-parametric filter widths And the number of iterationsThen, coarsening the image to obtain the first image in turnReformed image after sub-coarse granularity;

An extended and reorganized image acquisition module, for the firstReformed image obtained after sub-coarse granularityExpanding to obtain an expanded image;

The coincidence degree calculating module calculates the firstReformed image obtained after sub-coarse granularityIs a spread image of (a)And the firstReformed image obtained by sub-coarseningDegree of overlap between;

An image complexity calculating module for calculating image complexityAmplifying by exponential function

By combining all the technical schemes, the invention has the following beneficial effects:

In terms of image size limitation, the current technology requires that the processed images must be consistent in width and height, which limits the applicability of the method in common applications. The invention provides a complexity evaluation method capable of processing images with any aspect ratio, thereby expanding the application range of the technology and enabling the technology to be more suitable for the analysis requirement of common users on the complexity of conventional images.

In terms of complexity evaluation comprehensiveness, the existing method only considers the initial state and the final state of the image, and ignores information of each intermediate scale in the coarse graining process. The present invention proposes a new computational framework that incorporates images of all or more intermediate coarse-grain scales into the complexity computation to provide a more comprehensive and detailed image complexity description, thereby reflecting the true complexity of the image more accurately.

The complexity value generated by the traditional method is between 0 and 1 on the aspect of the result discrimination, so that the complexity difference between different images is not easy to identify. The invention aims to develop a new measurement standard or adjust the existing algorithm to increase the value range of the output complexity value, so that the complexity difference between different images can be distinguished more easily.

In terms of computational efficiency, conventional methods are inefficient in processing high resolution images or large data volume image sets. The invention provides a more efficient algorithm implementation or a more optimized calculation resource management scheme to improve the calculation efficiency of the coarse graining process and complexity evaluation, thereby being more suitable for the processing requirements of high-resolution images or large-scale image sets.

And aiming at the condition that the width and the height of the original images are inconsistent, adopting a filling strategy to adjust the images, so that all the processed images have uniform width and height, and ensuring the universality and the adaptability of the algorithm.

The invention not only considers the complexity of the final coarse-grained image, but also integrates the complexity changes of the images at each stage in the middle into the final complexity assessment, and the overall complexity is assessed by integrating the continuous changes. The complexity calculation efficiency is greatly improved by adjusting the numerical range of the complexity output (for example, from 0-1 to wider numerical values such as 21, 43 and 72), so that the degree of differentiation of the result is higher, the analysis and the interpretation are more convenient, and the algorithm optimization and the simplification of the calculation flow are realized. The method and the device can be applied to multiple fields of automatic classification, evaluation, search optimization and the like of the images by adopting multi-scale fusion to analyze the images and more accurately evaluate the complexity of the images.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and together with the description, serve to explain the principles of the disclosure;

FIG. 1 is a schematic diagram of an image complexity calculation method considering multi-scale fusion according to an embodiment of the present invention;

FIG. 2 is a simplified schematic illustration of an image expansion provided by an embodiment of the present invention;

FIG. 3 is a schematic diagram of an image complexity calculation system that takes into account multi-scale fusion according to an embodiment of the present invention;

FIG. 4 is a schematic illustration of an edge-filled black filling experiment according to an embodiment of the present invention;

FIG. 5 is a second experimental illustration of edge filling with black filling background according to an embodiment of the present invention;

FIG. 6 is a third experimental illustration of edge filling with black filling background provided by an embodiment of the present invention;

FIG. 7 is a fourth experimental illustration of edge filling with black filling background provided by an embodiment of the present invention;

FIG. 8 is a fifth experimental illustration of edge filling with black filling background provided by an embodiment of the present invention;

FIG. 9 is a sixth experimental illustration of edge filling with black filling background provided by an embodiment of the present invention;

in the figure, the method comprises a1 image background filling module, a2 iteration number value range determining module, a3 reforming image obtaining module, a 4 expansion reforming image obtaining module, a 5 contact ratio calculating module and a 6 image complexity calculating module.

Detailed Description

In order that the above objects, features and advantages of the invention will be readily understood, a more particular description of the invention will be rendered by reference to the appended drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The invention may be embodied in many other forms than described herein and similarly modified by those skilled in the art without departing from the spirit or scope of the invention, which is therefore not limited to the specific embodiments disclosed below.

The embodiment of the invention provides an image complexity calculation method, which realizes more accurate and efficient evaluation of image complexity through wide-high consistency adjustment, multi-scale complexity calculation, adjustment of a complexity output range and optimization of calculation efficiency. The method expands the numerical range of complexity output by comprehensively analyzing the changes of images with different sizes, optimizes the algorithm to improve the calculation efficiency, and is particularly suitable for processing large-scale image data. The image complexity calculating method and system considering multi-scale fusion provided by the embodiment of the invention have the innovation points that:

Embodiment 1 of the present invention provides an improved image complexity calculation method considering multi-scale fusion, which mainly includes:

(1) And adjusting the width and height consistency, namely adjusting the original image with any aspect ratio by using a filling or other image processing methods to adapt to the requirements of a complexity algorithm, and ensuring the width and height consistency of the image.

(2) Multi-scale complexity calculation, namely evaluating the overall image complexity by comprehensively analyzing the image changes of each size in the coarsening process instead of considering the original image and the picture coarsened for the last time.

(3) And (3) adjusting the complexity output range, namely innovatively expanding the complexity output numerical range, so that the result distinguishing degree is improved, and the complexity evaluation is finer and more practical.

(4) Optimization of computational efficiency by improving the algorithm and simplifying the computational process, improves the efficiency of complexity computation, especially when processing large-scale image data requiring a large amount of computation.

Embodiment 2, as a further detailed technical solution of the present invention, provides an image complexity calculating method considering multi-scale fusion according to an embodiment of the present invention, which specifically includes the following steps:

The background color of the general image filling is black, but since the invention needs to perform normalization processing (preprocessing) on the image data, the calculation formula of the normalization processing is as follows:

;

In the middle of Respectively the first image after pretreatmentLine 1The RGB channel normalized values for the column pixels,Respectively the first in the original imageLine 1RGB channel values for column pixels. Therefore, when the RGB channel value of the filling background is (128,128,128), the normalized value of the RGB channel after normalization of the filling background is (0, 0), so the invention performs classified discussion on the color of the channel background. In addition, the background filling mode is generally divided into edge filling and positioning filling (the original image is placed on the left side of the background filling area for comparison analysis), so the invention also carries out classified discussion on the filling mode of the channel background.

S2, determining widths of two super-parametric filtersAnd the number of iterationsCalculating the number of iterationsIs a value range of (a);

Specifically, the filter width is determined In general. Number of iterationsIs an superparameter in the coarsening process, and the width and height of the reformed image obtained by coarsening each time can be reducedThe minimum value is 1, so the invention innovatively provides that the value range of the iteration times N can be calculated according to the following formula:

;

The technical function of equation (4) is to determine the range of values of the number of iterations N.

S3, determining widths of two super-parametric filtersAnd the number of iterationsThen, coarsening the image to obtain the first image in turnReformed image after sub-coarse granularity;

Carry out the first step on the imagePerforming coarse graining treatment to obtain a product with a side length of the first sizeReformed image obtained after sub-coarse graining;

In the course of coarsening, the division of an image into multiple dimensions isThen adding the values of the pixels of the pixel block to obtain an average value to obtain a pixel point, and replacing the original pixel block with the pixel point. Traversing each pixel block, sequentially performing the operation, and finally finishing one coarse granularity to obtain an image with a side length of coarse granularityThe specific calculation formula is as follows:

;

First, theImages before sub-coarseningI.e. the firstReformed image obtained after sub-coarse grainingTherefore, the firstReformed image obtained after sub-coarse grainingIs the firstReformed image obtained after sub-coarse grainingA kind of electronic device;

Specifically, assume thatIs the firstThe reformed image obtained after the sub-coarse grain,Is the original image. The invention is applied to a multi-scale fusion technology when calculating the complexity of the image, and needs to calculateAnd (3) withThe overlap ratio between the two.

Due toIs of side length ofA kind of electronic deviceIn the process of calculationAnd (3) withIn the overlap ratio of (2), the dimensions are required to be uniform, and thus forPerforming expansion (nearest neighbor upsampling) to obtain the firstReformed image obtained by sub-coarseningIs a spread image of (a)As shown in the extended image in fig. 1, the calculation formula of the nearest neighbor upsampling extension is:

;

A simple calculation of the image expansion is shown in fig. 2, in which,Is the firstThe side length of the image obtained by the sub-coarse granularity is an integer,Is the side length of the expanded image and meets the following requirements,Is thatIs a picture of the expansion image of (a);

S5, calculating the first Reformed image obtained after sub-coarse granularityIs a spread image of (a)And the firstReformed image obtained by sub-coarseningDegree of overlap between;

Assume thatIs thatIs the first of (2)Line 1The pixel points are listed, so that it is a vector,And (3) withThe product of (2) is a vector product. The innovation of the invention proposes to calculate the firstReformed image obtained after sub-coarse granularityIs a spread image of (a)And the firstReformed image obtained by sub-coarseningDegree of overlap betweenThe calculation formula is as follows:

;

S6, calculating image complexityThe calculation formula is as follows:

;

In the formula, Is of complexityIs used for the control of the degree of freedom of the composition,Is the firstReformed image obtained by sub-coarseningIs a spread image of (a)And the firstReformed image obtained by sub-coarseningThe overlap ratio between the two.

Complexity of the imageObtained by exponential function amplificationThe expression is:

;

In the formula, In order to achieve an enlarged image complexity,For the complexity of the original image it is,Is the base of natural logarithms.

According to the embodiment, the image is adjusted by adopting the filling strategy aiming at the condition that the width and the height of the original image are inconsistent, so that all processed images have uniform width and height, and the universality and the adaptability of the algorithm are ensured.

The invention not only considers the complexity of the final coarse-grained image, but also integrates the complexity changes of the images at each stage in the middle into the final complexity assessment, and the overall complexity is assessed by integrating the continuous changes.

By adjusting the range of values of the complexity output (e.g., from 0-1 to a wider range of values such as 21, 43, 72), the discrimination of the results is higher, facilitating analysis and interpretation;

and the complexity calculation efficiency is greatly improved through algorithm optimization and calculation flow simplification.

Embodiment 3 as shown in fig. 3, an image complexity calculation system considering multi-scale fusion according to an embodiment of the present invention includes:

The image background filling module 1 is used for judging whether the pixel values of the width and the height of the input image are the same, if not, filling the image background, and if the larger value is odd, the image size +1 is changed into even;

the iteration number value range determining module 2 is used for determining the widths of two super-parameter filters And the number of iterationsCalculating the number of iterationsIs a value range of (a);

a reorganization image acquisition module 3 for determining two super-parametric filter widths And the number of iterationsThen, coarsening the image to obtain the first image in turnReformed image after sub-coarse granularity;

An extended and reorganized image acquisition module 4 for the firstReformed image obtained after sub-coarse granularityExpanding to obtain an expanded image;

The overlap ratio calculation module 5 calculates the firstReformed image obtained after sub-coarse granularityIs a spread image of (a)And the firstReformed image obtained by sub-coarseningDegree of overlap between;

An image complexity calculation module 6 for calculating image complexityAmplification is performed by an exponential function.

The technical improvement point can be applied to the following fields:

Petroleum leakage detection-image complexity can be used to detect the severity of petroleum leakage in sea water. By analyzing the image of the oil film in the seawater, the distribution and diffusion condition of the oil stain can be evaluated, so that corresponding cleaning measures can be adopted.

Magnetic field detection in which image complexity can be used to analyze the strength and distribution of a magnetic field. By visualizing the magnetic field lines, the complexity of the magnetic field can be more intuitively understood, which is important for physical experiment and engineering applications.

Medical imaging analysis in the field of medical imaging, image complexity may help doctors analyze the complexity of a lesion region, such as in MRI or CT scanning, complexity analysis may help identify and differentiate between different tissue types.

Remote sensing image analysis in the remote sensing field, image complexity can be used to analyze surface coverage types and changes, which is critical to environmental monitoring and urban planning.

According to the embodiment, the application range of the algorithm is widened, more kinds of images can be processed, the original aspect ratio of the images is not required to be limited, and the use convenience of a user is improved.

The invention provides finer and dynamic complexity assessment, which can reflect the change and characteristics of the image content more truly.

The invention improves the sensitivity and practicability of complexity evaluation, so that the complexity comparison among different images is more obvious, and the invention is convenient for further data processing and application.

The invention reduces the consumption of computing resources, accelerates the processing speed, has obvious advantages especially when processing a large number of or high-resolution images, and improves the user experience.

The image complexity under different conditions is compared according to the above procedure, the initial conditions α=2, k=11 are given, and the experimental legends of edge filling and black filling background are given, as shown in fig. 4-9. The Python version used in the invention is 3.8.0, the Pytorch-cuda version is 11.7, the GPU is GTX-1660-Super, the CPU is i5-12400F, and the comparative test analysis results are shown in Table 1:

TABLE 1 image complexity (C) comparison experiment results

Experimental results show that the MSE of the edge filling-gray is minimum, the average running time of the invention is 3945.44, the running time of the prior art is 31937.38s, 27991.94s are saved, and the efficiency is improved by about 87.65%.

(8) The image complexity results in table 1 are amplified in combination with equation (9), and the results are shown in table 2:

TABLE 2 image complexity (C _m) vs. experimental results

Experimental results show that the amplified data is more distinct and that the MES has the same law as before.

While the invention has been described with respect to what is presently considered to be the most practical and preferred embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but on the contrary, is intended to cover various modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

Claims

1. A method for calculating image complexity considering multi-scale fusion, characterized in that the method comprises:

S1, determine whether the pixel values of the width and height of the input image are the same. If they are not the same, fill the image background. If the larger value is an odd number, the image size is increased by 1 to become an even number;

S2, determines the width of two hyperparameter filters With the number of iterations , calculate the number of iterations The value range of

S3, performs coarse-grained processing on the image, and obtains the Renormalized image after sub-coarse granularity ;

S4, for The renormalized image obtained after the second coarse granularity Expand and get the extended image ;

S5, calculate the The renormalized image obtained after the second coarse granularity Extended image of With The renormalized image obtained by sub-coarse granularization The overlap between ;

S6, Calculate image complexity .

2. According to the image complexity calculation method considering multi-scale fusion as described in claim 1, it is characterized in that in step S1, after the image background is filled, the image data is normalized, and the background filling method is divided into edge filling and positioning filling. Positioning filling is to place the original image on the left side of the background filling area for comparative analysis, and the normalization calculation formula is:

;

In the formula, They are the first Row, No. The normalized values of the RGB channels of the column pixels, They are the first Row, No. RGB channel values of the column pixels.

3. The image complexity calculation method considering multi-scale fusion according to claim 1, characterized in that in step S2, the filter width , filter width With the number of iterations are two hyperparameters in the coarse-graining process. The width and height of the renormalized image obtained by each coarse-graining are reduced. times, and the minimum width and height is 1, the number of iterations The value range calculation formula is:

;

In the formula, For the The side length of the image obtained at the second coarse granularity is an integer, is the side length of the original image.

4. The method for calculating image complexity considering multi-scale fusion according to claim 3, characterized in that in step S3, the image is subjected to coarse-grained processing, comprising:

In the coarse-graining process, the image is divided into multiple sizes The pixel block is formed by adding the values of each pixel in the pixel block and taking the average value to obtain a pixel point, which is used to replace the original pixel block; each pixel block is traversed to complete a coarse-graining. The renormalized image obtained after the second coarse granularity , the calculation formula is:

;

In the formula, For the The renormalized image obtained by the second coarse-graining is Row, No. Column pixels; is the image before coarse-graining Row, No. The first pixel block Row, No. Column pixels; For the The renormalized image obtained by the second coarse-graining is Row, No. Column pixels.

5. The method for calculating image complexity considering multi-scale fusion according to claim 1, characterized in that in step S4, The renormalized image obtained after the second coarse granularity The nearest neighbor upsampling expansion is used to obtain the The renormalized image obtained by sub-coarse granularization Extended image of , the calculation formula for the nearest neighbor upsampling expansion is:

;

In the formula, for No. Row, No. Column pixels, for No. Row, No. Column pixels, for No. Row, No. Column pixels, for No. Row, No. Column pixels.

6. The image complexity calculation method considering multi-scale fusion according to claim 5 is characterized in that the renormalized image The side length is The renormalized image obtained after the second coarse granularity of times, expand the image and The size is consistent; and meets ;in, For the The side length of the image obtained at the second coarse granularity is an integer, is the side length of the expanded image, for The extended image of .

7. The method for calculating image complexity considering multi-scale fusion according to claim 1, characterized in that in step S5, The renormalized image obtained after the second coarse granularity Extended image of With The renormalized image obtained by sub-coarse granularization The overlap between ,include:

for No. Row, No. Column pixel points, a vector, and The product of is the vector product; the calculation formula is:

;

In the formula, For the The renormalized image obtained by -1 times of coarse granularity is Row, No. Column pixels.

8. The method for calculating image complexity considering multi-scale fusion according to claim 1, characterized in that in step S6, the image complexity The calculation formula is:

;

In the formula, is the component of complexity C, For the The renormalized image obtained by sub-coarse granularization Extended image of With The renormalized image obtained by sub-coarse granularization The overlap between .

9. The method for calculating image complexity considering multi-scale fusion according to claim 1, characterized in that the image complexity By exponentially magnifying , the expression is:

;

In the formula, is the complexity of the enlarged image, C is the complexity of the original image, is the base of natural logarithms.

10. A system for calculating image complexity considering multi-scale fusion, characterized in that the system implements the method for calculating image complexity considering multi-scale fusion as claimed in any one of claims 1 to 9, and the system comprises:

The image background filling module (1) is used to determine whether the pixel values of the width and height of the input image are the same. If they are not the same, the image background is filled. If the larger value is an odd number, the image size is increased by 1 to become an even number.

Iteration number range determination module (2) is used to determine the width of two hyperparameter filters With the number of iterations , calculate the number of iterations The value range of

The renormalized image is obtained by module (3), which is used to determine the two hyperparameter filter widths. With the number of iterations After that, the image is processed into coarse grains, and the Renormalized image after sub-coarse granularity ;

Expand the renormalized image acquisition module (4) to The renormalized image obtained after the second coarse granularity Expand and get the extended image ;

The overlap calculation module (5) calculates the The renormalized image obtained after the second coarse granularity Extended image of With The renormalized image obtained by sub-coarse granularization The overlap between ;

Image complexity calculation module (6), used to calculate image complexity , amplified by an exponential function.