CN113763291A - Performance evaluation method for preserving boundary filtering algorithm, intelligent terminal and storage medium - Google Patents
Performance evaluation method for preserving boundary filtering algorithm, intelligent terminal and storage medium Download PDFInfo
- Publication number
- CN113763291A CN113763291A CN202111031859.5A CN202111031859A CN113763291A CN 113763291 A CN113763291 A CN 113763291A CN 202111031859 A CN202111031859 A CN 202111031859A CN 113763291 A CN113763291 A CN 113763291A
- Authority
- CN
- China
- Prior art keywords
- pixel
- output image
- determining
- boundary
- input image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 85
- 238000001914 filtration Methods 0.000 title claims abstract description 81
- 238000011156 evaluation Methods 0.000 title claims abstract description 54
- 238000000034 method Methods 0.000 claims abstract description 38
- 230000014759 maintenance of location Effects 0.000 claims abstract description 12
- 238000004364 calculation method Methods 0.000 abstract description 13
- 230000000717 retained effect Effects 0.000 description 15
- 238000001514 detection method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 238000004590 computer program Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000002146 bilateral effect Effects 0.000 description 2
- 238000009792 diffusion process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a performance evaluation method for maintaining a boundary filtering algorithm, an intelligent terminal and a storage medium, wherein the method comprises the following steps: acquiring an input image, and determining an output image according to the input image and a to-be-evaluated retention boundary filtering algorithm; according to the input image and the output image, determining a global similarity parameter, a significant boundary similarity parameter and smoothness corresponding to the output image; and determining the performance of the boundary filtering algorithm to be evaluated and maintained according to the global similarity parameter, the significant boundary similarity parameter and the smoothness. The method evaluates the performance of the to-be-evaluated and maintained boundary filtering algorithm according to the global similarity parameter, the significant boundary similarity parameter and the smoothness, can achieve the same result as subjective evaluation, has accurate evaluation result, does not need human intervention, has high calculation speed, and can be suitable for any input image.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a performance evaluation method for a boundary filtering algorithm, an intelligent terminal and a storage medium.
Background
In many applications in the field of digital image processing, a boundary preserving filtering algorithm is used to smooth images, and for decades, various boundary preserving filters have been developed, and almost all filters have parameters for user configuration. Generally, a very experienced person can adjust the parameter value, and if the filtering result is found to be better by adjusting the parameter, the adjusted parameter is more appropriate; if the filtering result is found to be poor by adjusting the parameters, which indicates that the adjusted parameters are not suitable, the optimal values of the parameters of the boundary filter are maintained after a plurality of attempts. Therefore, it is important to evaluate the performance of the edge-preserving filter algorithm in order to configure optimal parameters for an arbitrary edge-preserving filter.
The conventional method for evaluating and maintaining the performance of the boundary filtering algorithm comprises an objective evaluation index and a subjective evaluation index, but the conventional objective evaluation index has larger performance evaluation error, and the subjective evaluation index requires that an input image comes from a reference database and needs human participation, so that the calculation speed is slow.
Thus, there is still a need for improvement and development of the prior art.
Disclosure of Invention
The invention provides a method for evaluating the performance of a boundary filter algorithm, an intelligent terminal and a storage medium, aiming at solving the technical problems that the existing method for evaluating the performance of the boundary filter algorithm is large in evaluation error, not suitable for any input image and low in calculation speed.
The technical scheme adopted by the invention for solving the problems is as follows:
in a first aspect, an embodiment of the present invention provides a method for evaluating performance of a retained boundary filtering algorithm, where the method includes:
acquiring an input image, and determining an output image according to the input image and a to-be-evaluated retention boundary filtering algorithm;
according to the input image and the output image, determining a global similarity parameter, a significant boundary similarity parameter and smoothness corresponding to the output image;
and determining the performance of the boundary filtering algorithm to be evaluated and maintained according to the global similarity parameter, the significant boundary similarity parameter and the smoothness.
The method for evaluating the performance of the boundary-preserving filter algorithm, wherein the step of determining the global similarity parameter, the significant boundary similarity parameter and the smoothness corresponding to the output image according to the input image and the output image comprises the following steps:
acquiring the number of pixel points corresponding to the input image, a first pixel value corresponding to each pixel point in the input image and a second pixel value corresponding to each pixel point in the output image;
and determining a global similarity parameter corresponding to the output image according to the number of the pixel points, the first pixel value and the second pixel value.
The method for evaluating performance of a retained boundary filtering algorithm includes the following steps after the step of obtaining the number of pixels corresponding to the input image, the first pixel value corresponding to each pixel point in the input image, and the second pixel value corresponding to each pixel point in the output image:
and determining the smoothness corresponding to the output image according to the number of the pixel points and the second pixel value.
The method for evaluating the performance of the boundary filtering algorithm includes the following steps:
according to the second pixel value, determining a first pixel gradient of each pixel point in the output image in the x direction and a second pixel gradient of each pixel point in the output image in the y direction;
and determining the smoothness corresponding to the output image according to the number of the pixel points, the first pixel gradient and the second pixel gradient.
The method for evaluating performance of a boundary-preserving filter algorithm, wherein the step of determining a global similarity parameter, a significant boundary similarity parameter and smoothness corresponding to the output image according to the input image and the output image further comprises:
determining a similarity score corresponding to the output image according to the input image and the output image;
and determining a salient boundary similarity parameter corresponding to the output image according to the similarity score.
The method for evaluating performance of a boundary-preserving filter algorithm, wherein the step of determining a similarity score corresponding to the output image from the input image and the output image comprises:
determining boundary position parameters corresponding to all pixel points in the input image according to the input image, and determining structural similarity parameters corresponding to all pixel points in the output image according to the input image and the output image;
and determining a similarity score corresponding to the output image according to the boundary position parameter and the structure similarity parameter.
The method for evaluating the performance of the boundary filtering algorithm includes the following steps of:
acquiring a first image block corresponding to each pixel point in the input image and a second image block corresponding to each pixel point in the output image, determining a first pixel average value and a first pixel variance corresponding to each pixel point in the input image according to the first image block, and determining a second pixel average value and a second pixel variance corresponding to each pixel point in the output image according to the second image block;
and determining a structural similarity parameter corresponding to each pixel point in the output image according to the first pixel average value, the first pixel variance, the second pixel average value and the second pixel variance.
The method for evaluating the performance of the retained boundary filtering algorithm, wherein the step of determining the performance of the retained boundary filtering algorithm to be evaluated according to the global similarity parameter, the significant boundary similarity parameter and the smoothness comprises the following steps:
determining an evaluation index corresponding to the to-be-evaluated retention boundary filtering algorithm according to the global similarity parameter, the significant boundary similarity parameter and the smoothness;
and determining the performance of the boundary filtering algorithm to be evaluated and maintained according to the evaluation index.
In a second aspect, an embodiment of the present invention further provides a performance evaluation apparatus for maintaining a boundary filtering algorithm, where the apparatus includes:
the image acquisition module is used for acquiring an input image and determining an output image according to the input image and a to-be-evaluated retention boundary filtering algorithm;
a parameter determining module, configured to determine, according to the input image and the output image, a global similarity parameter, a significant boundary similarity parameter, and smoothness corresponding to the output image;
and the performance evaluation module is used for determining the performance of the boundary filtering algorithm to be evaluated and kept according to the global similarity parameter, the significant boundary similarity parameter and the smoothness.
In a third aspect, an embodiment of the present invention provides an intelligent terminal, including a memory, and one or more programs, where the one or more programs are stored in the memory, and configured to be executed by one or more processors, where the one or more programs include steps for executing the performance evaluation method of the retained boundary filtering algorithm described in any one of the above.
In a fourth aspect, embodiments of the present invention provide a computer-readable storage medium, where instructions, when executed by a processor of an electronic device, enable the electronic device to perform the steps of the performance evaluation method for preserving a boundary filtering algorithm as described in any one of the above.
The invention has the beneficial effects that: according to the embodiment of the invention, an input image is firstly obtained, an output image is determined according to the input image and a to-be-evaluated retained boundary filtering algorithm, then a global similarity parameter, a significant boundary similarity parameter and smoothness corresponding to the output image are determined according to the input image and the output image, and finally the performance of the to-be-evaluated retained boundary filtering algorithm is determined according to the global similarity parameter, the significant boundary similarity parameter and the smoothness.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for evaluating performance of a preserving boundary filtering algorithm according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating a specific application of the SSIM algorithm based on the Sobel boundary mask according to an embodiment of the present invention when the input image is a color image;
FIG. 3 is a flowchart illustrating a specific application of the SSIM algorithm based on the Sobel boundary mask according to an embodiment of the present invention when the input image is a grayscale image;
FIG. 4 is a flowchart of an embodiment of a performance evaluation method for a preserving boundary filtering algorithm according to an embodiment of the present invention;
FIG. 5 is a schematic block diagram of a performance evaluation apparatus for preserving a boundary filtering algorithm according to an embodiment of the present invention;
fig. 6 is a schematic block diagram of an internal structure of an intelligent terminal according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer and clearer, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
It should be noted that, if directional indications (such as up, down, left, right, front, and back … …) are involved in the embodiment of the present invention, the directional indications are only used to explain the relative positional relationship between the components, the movement situation, and the like in a specific posture (as shown in the drawing), and if the specific posture is changed, the directional indications are changed accordingly.
In the field of digital image processing, many applications use a preserving boundary filtering algorithm to smooth an image, and typical examples include image denoising, fusion of different exposure images, high dynamic range image imaging, and application of preserving structure to remove texture. For decades, various kinds of hold boundary filters have emerged, such as anisotropic diffusion filters, bilateral filters, threshold transform filters, guided filters, filters based on wavelet transforms, and the like. Almost all of these filters have some parameters for user configuration. For example, the bilateral filter has two parameters σdAnd σrWherein σ isdControlling the geometric diffusion of the filter in the spatial domain, generally set according to the degree of low-pass filtering desired by the user, σrThe distribution of the luminance of the control filter in the image domain is typically set according to a user desired joint range of pixel values. In the actual filtering process, for different image contents and applications, it should be σdAnd σrDifferent values are configured. Generally, a very experienced person can adjust the parameter value, and if the filtering result is found to be better by adjusting the parameter, the adjusted parameter is more appropriate; if the filtering result is found to be poor by adjusting the parameters, which indicates that the adjusted parameters are not suitable, the optimal values of the parameters of the boundary filter are maintained after a plurality of attempts. Therefore, it is especially desirable to configure optimal parameters for arbitrary preserving boundary filters, and to evaluate the performance of preserving boundary filtering algorithmsIs of importance.
Methods for evaluating and maintaining the performance of the boundary filtering algorithm mainly fall into two categories: the method comprises objective evaluation indexes and subjective evaluation indexes, wherein common objective evaluation indexes comprise Mean Squared Error (MSE), Peak signal-to-noise ratio (PSNR) and Structural Similarity Index (SSIM), but when the performance of a boundary filtering algorithm is evaluated and maintained, the evaluation performance obtained by removing more parameters from texture information is worse, and the performance evaluation error is large. Recently, researchers have proposed evaluation indexes of weighted mean square error root (WRMSE) and Weighted Mean Absolute Error (WMAE), but the two evaluation indexes need to manually construct a reference database from which an input image must be derived, and in the evaluation process, professionals need to be invited to carry out subjective evaluation, so that the calculation speed is slow.
In order to solve the problems in the prior art, the embodiment provides a method for evaluating the performance of a retained boundary filtering algorithm, by which the performance of the retained boundary filtering algorithm can be accurately evaluated, and the method is fast in calculation speed and applicable to any input image because human intervention is not required. In specific implementation, an input image is firstly obtained, an output image is determined according to the input image and a to-be-evaluated retained boundary filtering algorithm, then a global similarity parameter, a significant boundary similarity parameter and smoothness corresponding to the output image are determined according to the input image and the output image, and finally the performance of the to-be-evaluated retained boundary filtering algorithm is determined according to the global similarity parameter, the significant boundary similarity parameter and the smoothness.
Exemplary method
The embodiment provides a performance evaluation method for a boundary filtering algorithm, which can be applied to an intelligent terminal. As shown in fig. 1 in particular, the method comprises:
and S100, acquiring an input image, and determining an output image according to the input image and a to-be-evaluated retention boundary filtering algorithm.
Specifically, the input image is an arbitrary image acquired by using an existing device with a photographing function, and the output image is an image obtained by processing the input image through a to-be-evaluated hold boundary filtering algorithm. In this embodiment, when performance evaluation needs to be performed on the retained boundary filtering algorithm, an input image is obtained through a device with a photographing function, and then the retained boundary filtering algorithm to be evaluated is used to process the input image to obtain an output image, so that performance evaluation is performed on the retained boundary filtering algorithm to be evaluated through the input image and the output image in the subsequent steps.
Step S200, according to the input image and the output image, determining a global similarity parameter, a significant boundary similarity parameter and smoothness corresponding to the output image.
Considering that the subjective evaluation method (i.e., observation by human eyes) for maintaining the boundary filtering algorithm is generally developed from the following three aspects, the global similarity between the input image and the output image, i.e., the similarity between the input image and the output image in contents, for example, there is no deviation in color, no deviation in spatial position, etc.; the output image has the characteristic of keeping the salient boundary of the input image, namely the output image keeps the most salient structure information in the input image; the smoothness of the output image itself, i.e. the smoothness as much as possible in the rest of the image, except for the significant structural information in the image. Therefore, in this embodiment, after an input image and an output image are acquired, according to the input image and the output image, a global similarity parameter, a significant boundary similarity parameter, and smoothness corresponding to the output image are determined, so that in a subsequent step, performance evaluation is performed on a to-be-evaluated retained boundary filtering algorithm according to the parameters. The global similarity parameter is used for evaluating the global similarity of the input image and the output image, and the smaller the global similarity parameter is, the more similar the input image and the output image are in the global situation; the salient boundary similarity parameter is used for evaluating the retentivity of the output image to the salient boundary of the input image, and the smaller the salient boundary similarity parameter is, the more similar the output image is to the input image in the most salient structure; the smoothness is used to evaluate the smoothness of the output image itself, and a smaller smoothness indicates a better smoothness of the output image itself.
In a specific embodiment, step S200 specifically includes:
step S210, acquiring the number of pixel points corresponding to the input image, a first pixel value corresponding to each pixel point in the input image and a second pixel value corresponding to each pixel point in the output image;
step S220, determining a global similarity parameter corresponding to the output image according to the number of the pixel points, the first pixel value, and the second pixel value.
Specifically, the number of the pixels refers to the number of the pixels in the input image, the output image is an image obtained by processing the input image through a hold boundary filtering algorithm, the number of the pixels also refers to the number of the pixels in the output image, the first pixel value is a pixel value corresponding to each pixel in the input image, and the second pixel value is a pixel value corresponding to each pixel in the output image.
In this embodiment, the distance between an input image and an output image in color is calculated to measure the global similarity between the input image and the output image, when determining the global similarity parameter corresponding to the output image, the number of pixels corresponding to the input image, a first pixel value corresponding to each pixel in the input image, and a second pixel value corresponding to each pixel in the output image are first obtained, and then the global similarity parameter corresponding to the output image is determined according to the number of pixels, the first pixel value, and the second pixel value. Wherein, the calculation formula of the global similarity parameter is as follows:
wherein D isc(Q, I) is global similarity parameter, I is input image, Q is output image, p is pixel point, IpIs a first pixel value, Q, corresponding to a pixel point p in the input imagepAnd M is the number of the pixel points, wherein M is the second pixel value corresponding to the pixel point p in the output image.
In a specific embodiment, after step S210, the method further includes:
and M220, determining the smoothness corresponding to the output image according to the number of the pixel points and the second pixel value.
Specifically, after the number of pixels corresponding to the input image, the first pixel value corresponding to each pixel in the input image, and the second pixel value corresponding to each pixel in the output image are obtained, the smoothness corresponding to the output image is determined according to the number of pixels and the second pixel value.
In one embodiment, step M220 specifically includes:
step M221, determining a first pixel gradient of each pixel point in the output image in the x direction and a second pixel gradient of each pixel point in the output image in the y direction according to the second pixel value;
and step M222, determining the smoothness corresponding to the output image according to the number of the pixel points, the first pixel gradient and the second pixel gradient.
Specifically, in this embodiment, when smoothness is determined according to the number of pixel points and a second pixel value, first, according to the second pixel value, a first pixel gradient of each pixel point in the output image in the x direction and a second pixel gradient of each pixel point in the output image in the y direction are determined, and then, according to the number of pixel points, the first pixel gradient and the second pixel gradient, smoothness corresponding to the output image is determined.
In one embodiment, the present embodiment calculates the smoothness by using the number of gradients (NOG), which is calculated by the formula
Wherein, p is a pixel point,
for a first pixel gradient in the x-direction of a pixel point p in the output image,
the number of pixels is the first pixel gradient of a pixel point p in the y direction in the output image, # represents the number of p satisfying the condition, and M is the number of pixels.
In another embodiment, the smoothness is calculated by using a Relative Total Variance (RTV), and first, according to the first pixel gradient and the second pixel gradient, a first relative variance of each pixel point in the output image in the x direction and a second relative variance of each pixel point in the output image in the y direction are determined, wherein the first relative variance and the second relative variance are calculated according to a formula
Wherein D isx(p) is the first relative variance, Dy(p) is the second relative variance, p and q are pixels in the output image, R (p) is a local region centered on pixel p, gp，qIs the weight of the pixel point q,
(xp，yp) Is the coordinate of pixel point p, (x)q，yq) Is the coordinate of the pixel point q, sigma is a constant,
for a first pixel gradient in the x-direction of a pixel point q in the output image,
is an output diagramAnd a second pixel gradient of the pixel point q in the image in the y direction. Then, according to the number of the pixel points, the first relative variance and the second relative variance, smoothness corresponding to the output image is determined, wherein a calculation formula of the smoothness is as follows:
wherein,
m is the number of pixel points, and epsilon is a smooth coefficient.
In a specific embodiment, step S200 further includes:
step S230, determining a similarity score corresponding to the output image according to the input image and the output image;
and step S240, determining a salient boundary similarity parameter corresponding to the output image according to the similarity score.
The effect of the boundary filtering is good and bad mainly in terms of whether the most significant boundary information of the image is stored and whether insignificant boundary information is filtered out, so that only the most significant boundary information in the image, not all boundary information, should be considered when evaluating the retentivity of the boundary filtering result to the input image boundary. In this embodiment, when determining the salient boundary similarity parameter corresponding to the output image, first, a similarity score corresponding to the output image is determined according to the input image and the output image, and then, according to the similarity score, the salient boundary similarity parameter corresponding to the output image is determined. The calculation formula of the significant boundary similarity parameter is as follows: ds(Q, I) ═ 1-MASKED-SSIM (Q, I), where D iss(Q, I) is the significant boundary similarity parameter and MASKED-SSIM (Q, I) is the similarity score.
In a specific embodiment, step S230 specifically includes:
step S231, determining boundary position parameters corresponding to each pixel point in the input image according to the input image, and determining structural similarity parameters corresponding to each pixel point in the output image according to the input image and the output image;
step S232, according to the boundary position parameter and the structure similarity parameter, determining a similarity score corresponding to the output image.
Specifically, the boundary position parameter is a pixel value of each pixel point of the input image after the boundary is extracted by the sobel boundary detection operator, for each pixel point in the input image, if the pixel point is located on the significant boundary, the pixel value of the pixel point after the boundary is extracted by the sobel boundary detection operator is 1, that is, the boundary position parameter of the pixel point is 1, and if the pixel point is not located on the significant boundary, the pixel value of the pixel point after the boundary is extracted by the sobel boundary detection operator is 0, that is, the boundary position parameter of the pixel point is 0. The Structural Similarity parameter is a Structural Similarity (SSIM) value corresponding to each pixel point in the output image, the SSIM value is often used for evaluating the Structural Similarity between the images, the Similarity score is obtained after the output image is subjected to an SSIM algorithm based on a Sobel boundary mask, and the larger the Similarity score is, the better the boundary of the output image which keeps the boundary filtering algorithm is to the input image is.
In this embodiment, when determining a similarity score corresponding to the output image, first, a sobel boundary detection operator is used to extract a boundary of the input image, so as to obtain boundary position parameters corresponding to each pixel point in the input image, where when the boundary of the input image is extracted using the sobel boundary detection operator, the boundary of the input image in each channel is extracted separately, for example, as shown in fig. 2 and 3, when the input image includes R, G, B channels, the boundary of the image in an R channel, a G channel, and a B channel is extracted separately using the sobel boundary detection operator; when the input image is a gray image, a boundary is extracted on one color channel. Then, the structural similarity parameter corresponding to each pixel point in the output image is determined according to the input image and the output image, similar to the boundary extraction, when the structural similarity parameter is determined, each channel of the input image is also determined, for example, as shown in fig. 2 and fig. 3, when the input image includes R, G, B channels, the structural similarity parameter corresponding to each pixel point in the output image on R, G, B channels is determined, and when the input image is a gray scale image, the structural similarity parameter corresponding to each pixel point in the output image on one channel is determined.
Obtaining boundary position parameters corresponding to each pixel point in an input image and structural similarity parameters corresponding to each pixel point in an output image, multiplying the boundary position parameters corresponding to each pixel point and the structural similarity parameters corresponding to each pixel point on corresponding channels respectively, and averaging non-zero values in multiplication results to obtain an average value, namely a similarity score corresponding to the output image.
In a specific embodiment, the determining, in step S231, the structural similarity parameter corresponding to each pixel point in the output image according to the input image and the output image specifically includes:
step S2311, obtaining a first image block corresponding to each pixel point in the input image and a second image block corresponding to each pixel point in the output image, determining a first pixel average value and a first pixel variance corresponding to each pixel point in the input image according to the first image block, and determining a second pixel average value and a second pixel variance corresponding to each pixel point in the output image according to the second image block;
step S2312, determining a structural similarity parameter corresponding to each pixel point in the output image according to the first pixel average value, the first pixel variance, the second pixel average value, and the second pixel variance.
Specifically, the first image block is an image block centered on each pixel point in the input image, the second image block is an image block centered on each pixel point in the output image, the first average pixel value is an average value of pixel values of each pixel point in the first image block, and the first variance pixel value is an average value of pixel values of each pixel point in the first image blockThe second pixel average value is an average value of pixel values of all pixel points in the second image block, and the second pixel variance is a variance of pixel values of all pixel points in the second image block. In this embodiment, when determining the structural similarity parameter corresponding to each pixel point in the output image, first obtain a first image block corresponding to each pixel point in the input image and a second image block corresponding to each pixel point in the output image, then determine a first pixel average value and a first pixel variance corresponding to each pixel point in the input image according to the first image block, determine a second pixel average value and a second pixel variance corresponding to each pixel point in the output image according to the second image block, and finally determine the structural similarity parameter corresponding to each pixel point in the output image according to the first pixel average value, the first pixel variance, the second pixel average value, and the second pixel variance. Wherein, the calculation formula of the structural similarity parameter is as follows:
wherein, X is a pixel point in the input image, Y is a pixel point in the output image, X is a first image block corresponding to the pixel point X in the input image, Y is a second image block corresponding to the pixel point Y in the output image, and μXIs the first pixel average value, mu, corresponding to the first image block XYIs the average value of the second pixels corresponding to the second image block Y,
for a first pixel variance corresponding to the first image block X,
a second pixel variance, σ, for the second image block YXYAs a covariance between the first image block X and the second image block Y, c1＝(k1L)2，c2＝(k2L)2，c1And c2For a constant to maintain stability, L is the dynamic range of the pixel, k1＝0.01，k2＝0.03。
Step S300, determining a global similarity parameter corresponding to the output image according to the number of the pixel points, the first pixel value and the second pixel value.
Specifically, after determining the global similarity parameter, the significant boundary similarity parameter, and the smoothness, the embodiment evaluates the performance of the to-be-evaluated preserving boundary filtering algorithm according to the global similarity parameter, the significant boundary similarity parameter, and the smoothness, and determines the performance of the to-be-evaluated preserving boundary filtering algorithm. In the embodiment, the performance of the boundary filtering algorithm to be evaluated is evaluated according to the global similarity parameter, the significant boundary similarity parameter and the smoothness, an objective evaluation index is constructed, the same result as the subjective evaluation can be achieved, and the calculation speed is high due to no need of human intervention, so that the method can be applied to a large database.
In one embodiment, step S300 specifically includes:
step S310, determining an evaluation index corresponding to the to-be-evaluated retention boundary filtering algorithm according to the global similarity parameter, the significant boundary similarity parameter and the smoothness;
and S320, determining the performance of the boundary filtering algorithm to be evaluated and maintained according to the evaluation index.
Specifically, in this embodiment, when the filtering algorithm of the retention boundary to be evaluated is evaluated according to the global similarity parameter, the significant boundary similarity parameter, and the smoothness, an evaluation index corresponding to the filtering algorithm of the retention boundary to be evaluated is determined according to the global similarity parameter, the significant boundary similarity parameter, and the smoothness. Due to the global similarity parameter Dc(Q, I), salient boundary similarity parameter DsThe (Q, I) and smoothness y (Q) differ in value range, requiring normalization of the three parameters to [0,1 ] prior to determination of the evaluation index]And then determining an evaluation index according to the normalized parameters, and determining the performance of the boundary filtering algorithm to be evaluated. Wherein, the calculation formula of the evaluation index is as follows:
wherein,
for the normalized global similarity parameter,
to be the normalized significant boundary similarity parameter,
is the normalized smoothness. For example, as shown in fig. 4, an input picture is input to the hold boundary filter algorithm G having parameters of parameter 1 and parameter 2 …, and an output image 1 and an output image 2 … are obtained, the evaluation indexes corresponding to the hold boundary filter algorithm G of the parameters are score 1 and score 2 …, and the parameter i corresponding to the score i having the smallest value is selected from the score 1 and score 2 …, and is determined as the optimum parameter of the hold boundary filter algorithm G.
Exemplary device
As shown in fig. 5, an embodiment of the present invention provides a performance evaluation apparatus for preserving a boundary filtering algorithm, including: an image acquisition module 510, a parameter determination module 520, and a performance evaluation module 530. Specifically, the image obtaining module 510 is configured to obtain an input image, and determine an output image according to the input image and a to-be-evaluated preserving boundary filtering algorithm. The parameter determining module 520 is configured to determine a global similarity parameter, a significant boundary similarity parameter, and a smoothness corresponding to the output image according to the input image and the output image. The performance evaluation module 530 is configured to determine the performance of the boundary filtering algorithm to be evaluated according to the global similarity parameter, the significant boundary similarity parameter, and the smoothness.
Based on the above embodiment, the present invention further provides an intelligent terminal, and a schematic block diagram thereof may be as shown in fig. 6. The intelligent terminal comprises a processor, a memory, a network interface, a display screen and a temperature sensor which are connected through a system bus. Wherein, the processor of the intelligent terminal is used for providing calculation and control capability. The memory of the intelligent terminal comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the intelligent terminal is used for being connected and communicated with an external terminal through a network. The computer program is executed by a processor to implement a performance evaluation method that preserves a boundary filtering algorithm. The display screen of the intelligent terminal can be a liquid crystal display screen or an electronic ink display screen, and the temperature sensor of the intelligent terminal is arranged inside the intelligent terminal in advance and used for detecting the operating temperature of internal equipment.
It will be understood by those skilled in the art that the block diagram shown in fig. 6 is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation to the intelligent terminal to which the solution of the present invention is applied, and a specific intelligent terminal may include more or less components than those shown in the figure, or combine some components, or have a different arrangement of components.
In one embodiment, an intelligent terminal is provided that includes a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for:
acquiring an input image, and determining an output image according to the input image and a to-be-evaluated retention boundary filtering algorithm;
according to the input image and the output image, determining a global similarity parameter, a significant boundary similarity parameter and smoothness corresponding to the output image;
and determining the performance of the boundary filtering algorithm to be evaluated and maintained according to the global similarity parameter, the significant boundary similarity parameter and the smoothness.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, databases, or other media used in embodiments provided herein may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
In summary, the present invention discloses a performance evaluation method for preserving a boundary filtering algorithm, an intelligent terminal and a storage medium, including: acquiring an input image, and determining an output image according to the input image and a to-be-evaluated retention boundary filtering algorithm; according to the input image and the output image, determining a global similarity parameter, a significant boundary similarity parameter and smoothness corresponding to the output image; and determining the performance of the boundary filtering algorithm to be evaluated and maintained according to the global similarity parameter, the significant boundary similarity parameter and the smoothness. The method evaluates the performance of the to-be-evaluated and maintained boundary filtering algorithm according to the global similarity parameter, the significant boundary similarity parameter and the smoothness, can achieve the same result as subjective evaluation, has accurate evaluation result, does not need human intervention, has high calculation speed, and can be suitable for any input image.
It is to be understood that the invention is not limited to the examples described above, but that modifications and variations may be effected thereto by those of ordinary skill in the art in light of the foregoing description, and that all such modifications and variations are intended to be within the scope of the invention as defined by the appended claims.
Claims (10)
1. A performance evaluation method for a boundary-preserving filter algorithm is characterized by comprising the following steps:
acquiring an input image, and determining an output image according to the input image and a to-be-evaluated retention boundary filtering algorithm;
according to the input image and the output image, determining a global similarity parameter, a significant boundary similarity parameter and smoothness corresponding to the output image;
and determining the performance of the boundary filtering algorithm to be evaluated and maintained according to the global similarity parameter, the significant boundary similarity parameter and the smoothness.
2. The method of claim 1, wherein the step of determining the global similarity parameter, the significant boundary similarity parameter, and the smoothness corresponding to the output image according to the input image and the output image comprises:
acquiring the number of pixel points corresponding to the input image, a first pixel value corresponding to each pixel point in the input image and a second pixel value corresponding to each pixel point in the output image;
and determining a global similarity parameter corresponding to the output image according to the number of the pixel points, the first pixel value and the second pixel value.
3. The method for evaluating performance of a preserving boundary filtering algorithm according to claim 2, wherein the step of obtaining the number of pixels corresponding to the input image, the first pixel value corresponding to each pixel in the input image, and the second pixel value corresponding to each pixel in the output image further comprises:
and determining the smoothness corresponding to the output image according to the number of the pixel points and the second pixel value.
4. The method of claim 3, wherein the step of determining the smoothness corresponding to the output image according to the number of pixels and the second pixel value comprises:
according to the second pixel value, determining a first pixel gradient of each pixel point in the output image in the x direction and a second pixel gradient of each pixel point in the output image in the y direction;
and determining the smoothness corresponding to the output image according to the number of the pixel points, the first pixel gradient and the second pixel gradient.
5. The method of claim 2, wherein the step of determining the global similarity parameter, the significant boundary similarity parameter, and the smoothness corresponding to the output image according to the input image and the output image further comprises:
determining a similarity score corresponding to the output image according to the input image and the output image;
and determining a salient boundary similarity parameter corresponding to the output image according to the similarity score.
6. The method of claim 5, wherein the step of determining the similarity score corresponding to the output image according to the input image and the output image comprises:
determining boundary position parameters corresponding to all pixel points in the input image according to the input image, and determining structural similarity parameters corresponding to all pixel points in the output image according to the input image and the output image;
and determining a similarity score corresponding to the output image according to the boundary position parameter and the structure similarity parameter.
7. The method of claim 6, wherein the step of determining the structural similarity parameter corresponding to each pixel in the output image according to the input image and the output image comprises:
acquiring a first image block corresponding to each pixel point in the input image and a second image block corresponding to each pixel point in the output image, determining a first pixel average value and a first pixel variance corresponding to each pixel point in the input image according to the first image block, and determining a second pixel average value and a second pixel variance corresponding to each pixel point in the output image according to the second image block;
and determining a structural similarity parameter corresponding to each pixel point in the output image according to the first pixel average value, the first pixel variance, the second pixel average value and the second pixel variance.
8. The method of claim 1, wherein the step of determining the performance of the boundary filtering algorithm to be evaluated according to the global similarity parameter, the significant boundary similarity parameter and the smoothness comprises:
determining an evaluation index corresponding to the to-be-evaluated retention boundary filtering algorithm according to the global similarity parameter, the significant boundary similarity parameter and the smoothness;
and determining the performance of the boundary filtering algorithm to be evaluated and maintained according to the evaluation index.
9. An intelligent terminal comprising a memory, and one or more programs, wherein the one or more programs are stored in the memory, and wherein the one or more programs configured to be executed by the one or more processors comprise steps for performing a method of performance evaluation that preserves boundary filtering algorithms as claimed in any of claims 1-8.
10. A computer readable storage medium having instructions which, when executed by a processor of an electronic device, enable the electronic device to perform the steps of the method of performance evaluation that preserves boundary filtering algorithms according to any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111031859.5A CN113763291B (en) | 2021-09-03 | 2021-09-03 | Performance evaluation method for maintaining boundary filtering algorithm, intelligent terminal and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111031859.5A CN113763291B (en) | 2021-09-03 | 2021-09-03 | Performance evaluation method for maintaining boundary filtering algorithm, intelligent terminal and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113763291A true CN113763291A (en) | 2021-12-07 |
| CN113763291B CN113763291B (en) | 2023-08-29 |
Family
ID=78792846
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111031859.5A Active CN113763291B (en) | 2021-09-03 | 2021-09-03 | Performance evaluation method for maintaining boundary filtering algorithm, intelligent terminal and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113763291B (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3392307B2 (en) * | 1995-11-02 | 2003-03-31 | 松下電器産業株式会社 | Image signal smoothing apparatus and image signal smoothing method |
| CN101996406A (en) * | 2010-11-03 | 2011-03-30 | 中国科学院光电技术研究所 | No-reference structure definition image quality evaluation method |
| CN102265618A (en) * | 2008-12-25 | 2011-11-30 | 夏普株式会社 | Image decoding device and image coding device |
| US20160284065A1 (en) * | 2015-03-24 | 2016-09-29 | Intel Corporation | Non-local means image denoising with an adaptive directional spatial filter |
| CN106920232A (en) * | 2017-02-22 | 2017-07-04 | 武汉大学 | Gradient similarity graph image quality evaluation method and system based on conspicuousness detection |
| CN107194902A (en) * | 2017-06-09 | 2017-09-22 | 西安电子科技大学 | The method that wave filter parameter is automatically determined in image co-registration |
| CN108052078A (en) * | 2017-12-11 | 2018-05-18 | 深圳信息职业技术学院 | A kind of tele-control system of building castoff processing equipment |
| CN109146892A (en) * | 2018-07-23 | 2019-01-04 | 北京邮电大学 | A kind of image cropping method and device based on aesthetics |
| CN109658370A (en) * | 2018-11-29 | 2019-04-19 | 天津大学 | Image split-joint method based on mixing transformation |
| CN111784620A (en) * | 2020-07-06 | 2020-10-16 | 太原科技大学 | Light field camera full-focus image fusion algorithm for guiding angle information by spatial information |
| CN112837319A (en) * | 2021-03-29 | 2021-05-25 | 深圳大学 | Intelligent evaluation method, device, equipment and medium for real distorted image quality |
| CN113220919A (en) * | 2021-05-17 | 2021-08-06 | 河海大学 | Dam defect image text cross-modal retrieval method and model |
-
2021
- 2021-09-03 CN CN202111031859.5A patent/CN113763291B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3392307B2 (en) * | 1995-11-02 | 2003-03-31 | 松下電器産業株式会社 | Image signal smoothing apparatus and image signal smoothing method |
| CN102265618A (en) * | 2008-12-25 | 2011-11-30 | 夏普株式会社 | Image decoding device and image coding device |
| CN101996406A (en) * | 2010-11-03 | 2011-03-30 | 中国科学院光电技术研究所 | No-reference structure definition image quality evaluation method |
| US20160284065A1 (en) * | 2015-03-24 | 2016-09-29 | Intel Corporation | Non-local means image denoising with an adaptive directional spatial filter |
| US9576346B2 (en) * | 2015-03-24 | 2017-02-21 | Intel Corporation | Non-local means image denoising with an adaptive directional spatial filter |
| CN106920232A (en) * | 2017-02-22 | 2017-07-04 | 武汉大学 | Gradient similarity graph image quality evaluation method and system based on conspicuousness detection |
| CN107194902A (en) * | 2017-06-09 | 2017-09-22 | 西安电子科技大学 | The method that wave filter parameter is automatically determined in image co-registration |
| CN108052078A (en) * | 2017-12-11 | 2018-05-18 | 深圳信息职业技术学院 | A kind of tele-control system of building castoff processing equipment |
| CN109146892A (en) * | 2018-07-23 | 2019-01-04 | 北京邮电大学 | A kind of image cropping method and device based on aesthetics |
| CN109658370A (en) * | 2018-11-29 | 2019-04-19 | 天津大学 | Image split-joint method based on mixing transformation |
| CN111784620A (en) * | 2020-07-06 | 2020-10-16 | 太原科技大学 | Light field camera full-focus image fusion algorithm for guiding angle information by spatial information |
| CN112837319A (en) * | 2021-03-29 | 2021-05-25 | 深圳大学 | Intelligent evaluation method, device, equipment and medium for real distorted image quality |
| CN113220919A (en) * | 2021-05-17 | 2021-08-06 | 河海大学 | Dam defect image text cross-modal retrieval method and model |
Non-Patent Citations (4)
| Title |
|---|
| FARBMAN Z: "Edge-preserving decompositions for multi-scale tone and detail manipulation", 《ACM TRANS. GRAPHICS (TOG)》 * |
| HUI YIN: "Adaptive convolutional neural network for large change in video object segmentation", 《IET COMPUTER VISION》 * |
| LIANG Z: "A benchmark for edge-preserving image smoothing", 《IEEE TRANS. IMAGE PROCESS. (TIP)》 * |
| RAHARDJA S: "Detail-enhanced exposure fusion", 《IEEE TRANS IMAGE PROCESS》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113763291B (en) | 2023-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7324701B2 (en) | Image noise reduction | |
| CN108109147B (en) | No-reference quality evaluation method for blurred image | |
| US20060291741A1 (en) | Image processing apparatus, image processing method, program, and recording medium therefor | |
| US20090136152A1 (en) | Image correcting apparatus and method, and image correcting program, and look-up table creating apparatus and method, and look-up table creating program | |
| CN110232670B (en) | Method for enhancing visual effect of image based on high-low frequency separation | |
| US20090161983A1 (en) | Method for automatic detection and correction of halo artifacts in images | |
| WO2017100971A1 (en) | Deblurring method and device for out-of-focus blurred image | |
| KR102582261B1 (en) | Method for determining a point spread function of an imaging system | |
| DE102010001520A1 (en) | Iris detection system and method supported by an aircraft sensor | |
| CN111368717A (en) | Sight line determining method and device, electronic equipment and computer readable storage medium | |
| CN110378182A (en) | Image analysis apparatus, method for analyzing image and recording medium | |
| CN113327206B (en) | Image fuzzy processing method of intelligent power transmission line inspection system based on artificial intelligence | |
| CN113222866A (en) | Gray scale image enhancement method, computer readable medium and computer system | |
| CN112419270B (en) | No-reference image quality evaluation method and device under meta-learning and computer equipment | |
| CN112132925A (en) | Method and device for reconstructing underwater image color | |
| CN111862040A (en) | Portrait picture quality evaluation method, device, equipment and storage medium | |
| DE102016125758B4 (en) | Systems and methods for estimating the modulation transfer function in an optical system | |
| CN113763291B (en) | Performance evaluation method for maintaining boundary filtering algorithm, intelligent terminal and storage medium | |
| CN111951295A (en) | Method and device for determining flight trajectory based on polynomial fitting high precision and electronic equipment | |
| CN111951178A (en) | Image processing method and device for remarkably improving image quality and electronic equipment | |
| Imamura et al. | Image denoising using non-local means for Poisson noise | |
| CN112825189A (en) | Image defogging method and related equipment | |
| TWI783721B (en) | Method for determining deblur filter coefficients | |
| CN114972084A (en) | Image focusing accuracy evaluation method and system | |
| CN111354048B (en) | Quality evaluation method and device for obtaining pictures by facing camera |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |