1. An image denoising method based on heterogeneous noise characteristics is characterized by comprising the following steps:

converting a three-dimensional data set corresponding to an initial image into a two-dimensional data set, and decomposing the two-dimensional data set after normalization operation to obtain the initial data set, wherein the initial data set comprises a first matrix and a second matrix;

calculating a probability density distribution function according to the initial data set, and establishing a decomposition model obeying each pixel point in the target data set according to the probability density distribution function;

optimizing the decomposition model by using a variational Bayesian method to obtain a parameter formula corresponding to each parameter of the decomposition model;

iterating all the parameter formulas to obtain the value of each parameter in the decomposition model;

calculating a vector corresponding to the first matrix and a vector corresponding to the second matrix according to the values of the parameters in the decomposition model to obtain a target image;

judging whether the quality of the target image meets the standard or not;

if the quality of the target image meets the standard, outputting the target image;

and if the quality of the target image does not meet the standard, continuing to iterate until the quality of the target image meets the standard.

2. The method of claim 1, wherein the probability density distribution function isz_ij～IG(a₀，b₀) Wherein x is_ijFor each pixel point of said initial data set, u_iAnd v_jThe ith row and jth column vectors, z, of the first matrix and the second matrix, respectively_ijIs the variance of a Gaussian distribution, a₀And b₀Is a preset constant.

3. The method of claim 2, wherein the decomposition model is expressed asτ_u～G(c₀，d₀)，τ_v～G(e₀，f₀) Wherein, τ_uIs u_iA priori distribution of_vIs v is_jA priori distribution of c₀、d₀、e₀And f₀Are all preset constants.

4. The method according to claim 3, wherein the step of optimizing the decomposition model by using a variational Bayesian method to obtain a parameter formula corresponding to each parameter of the decomposition model comprises:

calculating posterior distribution of each parameter of the decomposition model according to the variational Bayes method;

respectively calculating an iterative formula of each parameter of the decomposition model according to the posterior distribution of each parameter of the decomposition model;

and taking the iterative formula as a parameter formula corresponding to each parameter of the decomposition model.

5. The method of claim 4, wherein u is the decomposition model parameter_iA posterior distribution of (a) and (v)_jThe posterior distribution of (2) is that the posterior distributions of the Gaussian distribution are all Gaussian distributions;

τ_ua posterior distribution of (d) and_vthe posterior distribution of (a) is a gamma distribution;

z_ijthe posterior distribution of (a) is an inverse gamma distribution.

6. The method of claim 1, wherein the step of determining whether the quality of the target image meets a criterion comprises:

judging whether the root mean square error of the target image is smaller than or equal to a threshold value;

if the root mean square error of the target image is smaller than or equal to the threshold, judging that the quality of the target image meets the standard;

and if the root mean square error of the target image is larger than the threshold value, judging that the quality of the target image does not meet the standard.

7. An electronic device, characterized in that the electronic device comprises:

at least one processor; and the number of the first and second groups,

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of image denoising based on heterogeneous noise characteristics of any of the preceding claims 1-6.

8. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the heterogeneous noise characteristics-based image denoising method of any one of claims 1-6.

Background

At present, image preprocessing plays an important role in digital image processing, and the quality of an image directly influences the classification, identification and segmentation of the image. The image denoising is a basic flow of image preprocessing, and the quality of an image denoising effect directly influences the subsequent processing process of an image. With the development of science and technology and the coming of the big data era, image denoising does not only process a certain image, but needs to process a whole set of image data set. The existing image denoising method extracts a low-rank subspace of a data set according to the similarity of images in the data set, the denoising effect is poor, or the noise in the data set is described by using mixed Gaussian distribution and mixed exponential power distribution, so that the calculated amount is huge.

Therefore, the existing image denoising method has the problems of poor calculation efficiency and poor denoising effect.

Disclosure of Invention

In view of this, embodiments of the present disclosure provide an image denoising method, device and medium based on heterogeneous noise characteristics, which at least partially solve the problems of poor computational efficiency and denoising effect in the prior art.

In a first aspect, an embodiment of the present disclosure provides an image denoising method based on heterogeneous noise characteristics, including:

converting a three-dimensional data set corresponding to an initial image into a two-dimensional data set, and decomposing the two-dimensional data set after normalization operation to obtain the initial data set, wherein the initial data set comprises a first matrix and a second matrix;

calculating a probability density distribution function according to the initial data set, and establishing a decomposition model obeying each pixel point in the target data set according to the probability density distribution function;

optimizing the decomposition model by using a variational Bayesian method to obtain a parameter formula corresponding to each parameter of the decomposition model;

iterating all the parameter formulas to obtain the value of each parameter in the decomposition model;

calculating a vector corresponding to the first matrix and a vector corresponding to the second matrix according to the values of the parameters in the decomposition model to obtain a target image;

judging whether the quality of the target image meets the standard or not;

if the quality of the target image meets the standard, outputting the target image;

and if the quality of the target image does not meet the standard, continuing to iterate until the quality of the target image meets the standard.

According to a specific implementation of the embodiments of the present disclosure, the probability density distribution function isz_ij～IG(a₀，b₀) Wherein x is_ijFor each pixel point of said initial data set, u_iAnd v_jThe ith row and jth column vectors, z, of the first matrix and the second matrix, respectively_ijIs the variance of a Gaussian distribution, a₀And b₀Is a preset constant.

According to a specific implementation manner of the embodiment of the present disclosure, the expression of the decomposition model isτ_i～G(c₀，d₀)，τ_v～G(e₀，f₀) Wherein, τ_uIs u_iA priori distribution of_vIs v is_jA priori distribution of c₀、d₀、e₀And f₀Are all preset constants.

According to a specific implementation manner of the embodiment of the present disclosure, the step of optimizing the decomposition model by using a variational bayesian method to obtain a parameter formula corresponding to each parameter of the decomposition model includes:

calculating posterior distribution of each parameter of the decomposition model according to the variational Bayes method;

respectively calculating an iterative formula of each parameter of the decomposition model according to the posterior distribution of each parameter of the decomposition model;

and taking the iterative formula as a parameter formula corresponding to each parameter of the decomposition model.

According to a specific implementation manner of the embodiment of the present disclosure, u in each parameter of the decomposition model_iA posterior distribution of (a) and (v)_jThe posterior distribution of (2) is that the posterior distributions of the Gaussian distribution are all Gaussian distributions;

τ_ua posterior distribution of (d) and_vthe posterior distribution of (a) is a gamma distribution;

z_ijthe posterior distribution of (a) is an inverse gamma distribution.

According to a specific implementation manner of the embodiment of the present disclosure, the step of determining whether the quality of the target image meets a standard includes:

judging whether the root mean square error of the target image is smaller than or equal to a threshold value;

if the root mean square error of the target image is smaller than or equal to the threshold, judging that the quality of the target image meets the standard;

and if the root mean square error of the target image is larger than the threshold value, judging that the quality of the target image does not meet the standard.

In a second aspect, an embodiment of the present disclosure further provides an electronic device, including:

at least one processor; and the number of the first and second groups,

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method for image denoising based on heterogeneous noise characteristics of the first aspect or any implementation of the first aspect.

In a third aspect, the disclosed embodiments also provide a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the heterogeneous noise characteristic-based image denoising method in the first aspect or any implementation manner of the first aspect.

In a fourth aspect, the present disclosure also provides a computer program product including a computer program stored on a non-transitory computer readable storage medium, the computer program including program instructions which, when executed by a computer, cause the computer to execute the heterogeneous noise characteristic-based image denoising method in the first aspect or any implementation manner of the first aspect.

The image denoising scheme based on the heterogeneous noise characteristics in the embodiment of the disclosure includes: converting a three-dimensional data set corresponding to an initial image into a two-dimensional data set, and decomposing the two-dimensional data set after normalization operation to obtain the initial data set, wherein the initial data set comprises a first matrix and a second matrix; calculating a probability density distribution function according to the initial data set, and establishing a decomposition model obeying each pixel point in the target data set according to the probability density distribution function; optimizing the decomposition model by using a variational Bayesian method to obtain a parameter formula corresponding to each parameter of the decomposition model; iterating all the parameter formulas to obtain the value of each parameter in the decomposition model; calculating a vector corresponding to the first matrix and a vector corresponding to the second matrix according to the values of the parameters in the decomposition model to obtain a target image; judging whether the quality of the target image meets the standard or not; if the quality of the target image meets the standard, outputting the target image; and if the quality of the target image does not meet the standard, continuing to iterate until the quality of the target image meets the standard.

The beneficial effects of the embodiment of the disclosure are: by the scheme, the image data are preprocessed, the decomposition model which obeys each pixel point in the target data set is established, the denoised target image is output through the value of each parameter in the iterative decomposition model, and the calculation efficiency and the denoising effect are improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings needed to be used in the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present disclosure, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a schematic flowchart of an image denoising method based on heterogeneous noise characteristics according to an embodiment of the present disclosure;

fig. 2 is a schematic diagram of a decomposition model involved in an image denoising method based on heterogeneous noise characteristics according to an embodiment of the present disclosure;

fig. 3 is a schematic diagram illustrating comparison of calculation results of algorithms involved in an image denoising method based on heterogeneous noise characteristics according to an embodiment of the present disclosure;

fig. 4 is a schematic diagram illustrating comparison of denoising effects of algorithms related to an image denoising method based on heterogeneous noise characteristics according to an embodiment of the present disclosure;

fig. 5 is a schematic diagram illustrating comparison of denoising effects of algorithms related to another image denoising method based on heterogeneous noise characteristics according to an embodiment of the present disclosure;

fig. 6 is a schematic diagram of an electronic device provided in an embodiment of the present disclosure.

Detailed Description

The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

The embodiments of the present disclosure are described below with specific examples, and other advantages and effects of the present disclosure will be readily apparent to those skilled in the art from the disclosure in the specification. It is to be understood that the described embodiments are merely illustrative of some, and not restrictive, of the embodiments of the disclosure. The disclosure may be embodied or carried out in various other specific embodiments, and various modifications and changes may be made in the details within the description without departing from the spirit of the disclosure. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict. All other embodiments, which can be derived by a person skilled in the art from the embodiments disclosed herein without making any creative effort, shall fall within the protection scope of the present disclosure.

It is noted that various aspects of the embodiments are described below within the scope of the appended claims. It should be apparent that the aspects described herein may be embodied in a wide variety of forms and that any specific structure and/or function described herein is merely illustrative. Based on the disclosure, one skilled in the art should appreciate that one aspect described herein may be implemented independently of any other aspects and that two or more of these aspects may be combined in various ways. For example, an apparatus may be implemented and/or a method practiced using any number of the aspects set forth herein. Additionally, such an apparatus may be implemented and/or such a method may be practiced using other structure and/or functionality in addition to one or more of the aspects set forth herein.

It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present disclosure, and the drawings only show the components related to the present disclosure rather than the number, shape and size of the components in actual implementation, and the type, amount and ratio of the components in actual implementation may be changed arbitrarily, and the layout of the components may be more complicated.

In addition, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects may be practiced without these specific details.

At present, image preprocessing plays an important role in digital image processing, and the quality of an image directly influences the classification, identification and segmentation of the image. The image denoising is a basic flow of image preprocessing, and the quality of an image denoising effect directly influences the subsequent processing process of an image. With the development of science and technology and the coming of the big data era, image denoising does not only process a certain image, but needs to process a whole set of image data set. Therefore, the low rank matrix decomposition algorithm suitable for processing large data sets is more and more emphasized by scholars. The main principle is that according to the similarity of images in a data set, a low-rank subspace of the data set is extracted, and therefore noise in the data set is removed quickly and efficiently.

The modeling methods of the low-rank matrix decomposition of the mainstream can be classified into two major categories, that is, the low-rank matrix decomposition based on the L1 norm model and the low-rank matrix decomposition based on the L2 norm model, according to the model classification. The low-rank matrix decomposition method of the L1 norm model is mainly based on Laplace noise distribution to establish mathematical model solution, and the representative algorithm comprises RegL1ALM, VBMFL1 and the like, while the low-rank matrix decomposition method of the L2 norm model is mainly based on Gaussian noise distribution to establish mathematical model solution, and the representative algorithm comprises L2 Wiberg. The L1 norm model is more robust to outliers than the L2 norm model. And the computation speed of the L2 model is faster compared with the L1 norm model.

However, whether the model is an L1 norm model or an L2 norm model, the noise needs to be described mathematically. When noise can be expressed well by a mathematical formula, the model can show excellent effect. To characterize the noise in the data set, the student first selected the most common gaussian distribution to characterize the noise doping in the data set, typically represented by the algorithm L2 Wiberg. However, the model that uses gaussian distribution to characterize noise can only exert the best effect in the data set containing gaussian noise, but the effect is not ideal for the data set with more outliers. Since then, the scholars have focused on the laplacian distribution with thick tail features, and the abnormal points can be better processed by using the laplacian distribution, and typical representative algorithms are VBMFL1 and PRMF. However, the noise in nature is very complex, and it is difficult to delineate the noise in the data set using only one probability distribution. Therefore, scholars have proposed characterizing noise based on a mixed gaussian distribution and a mixed exponential power distribution, which represent MOG and PMOEP, respectively. Although the mixed gaussian distribution and the mixed exponential power distribution can better depict complex noise in the data set, the complexity of the algorithm is increased, and the method needs to be operated for a long time on computers with poor performance.

From the development process of the low rank matrix decomposition algorithm, scholars characterize noise in a data set by using more and more complex probability density functions, although the accuracy of algorithm processing is better and better, and meanwhile, longer operation time is required. In addition, the algorithms all use the same probability density distribution to characterize the noise of each pixel point in the data set, so that the uniqueness of the noise is ignored, and the model needs to be trained more times to adjust the parameters of the model to achieve the overall optimal effect. Therefore, considering the difference of noise on each pixel point may improve the calculation accuracy of the model and reduce the calculation time of the model.

The embodiment of the disclosure provides an image denoising method based on heterogeneous noise characteristics, which can be applied to a denoising process of heterogeneous noise of an electronic image.

Referring to fig. 1, a schematic flow chart of an image denoising method based on heterogeneous noise characteristics according to an embodiment of the present disclosure is provided. As shown in fig. 1, the method mainly comprises the following steps:

s101, converting a three-dimensional data set corresponding to an initial image into a two-dimensional data set, and decomposing the two-dimensional data set after normalization operation to obtain the initial data set, wherein the initial data set comprises a first matrix and a second matrix;

in specific implementation, considering that a data set corresponding to the initial image is generally three-dimensional and is inconvenient for calculation processing, the three-dimensional data set corresponding to the initial image may be converted into a two-dimensional data set, the two-dimensional data set is subjected to normalization operation, the two-dimensional data set after the normalization operation is a high-dimensional matrix, the two-dimensional data set may be processed, the two-dimensional data set is decomposed into two low-dimensional matrices by the high-dimensional matrix to form the initial data set, and the influence of an abnormal point on an algorithm is reduced, so that an error in a subsequent calculation process is smaller, the initial data set may include a first matrix of all row vectors in the initial image and a second matrix of all column vectors in the initial image, and the first matrix and the second matrix are both low-dimensional matrices.

S102, calculating a probability density distribution function according to the initial data set, and establishing a decomposition model obeying each pixel point in the target data set according to the probability density distribution function;

optionally, the probability density distribution function isz_ij～IG(a₀，b₀) Wherein x is_ijFor each pixel point of said initial data set, u_iAnd v_jThe ith row and jth column vectors, z, of the first matrix and the second matrix, respectively_ijIs the variance of a Gaussian distribution, a₀And b₀Is a preset constant.

Further, the expression of the decomposition model isτ_u～G(c₀，d₀)，τ_v～G(e₀，f₀) Wherein, τ_uIs u_iA priori distribution of_vIs v is_jA priori distribution of c₀、d₀、e₀And f₀Are all preset constants.

In particular, a data size of m × n may be given,where m is the dimension of the data matrix, n is the number of data, and each column vector contains all the information of one picture. The low rank matrix decomposition problem can be expressed as an optimization problem

WhereinAndis a low dimensional matrix (r < min (m, n)). W is the deletion matrix if x_ijIs lost, then w_ij0, otherwise w_ij1. As indicates the hadamard product.Represents L_pAnd (4) norm.

Equation (1-1) is equivalent to under the framework of maximum likelihood estimation

Wherein x is_ijFor each pixel point of said initial data set, u_iAnd v_jThe ith row and jth column vectors, epsilon, of said first matrix and said second matrix, respectively_ijRepresenting a data set x_ijThe noise in the interior.

Assuming that the noise of each pixel point satisfies different Gaussian distributions

∈_ij～N(0，z_ij)，z_ij～IG(a₀，b₀) (1-3)

Wherein N (0, z)_ij) Representing a mean of 0 and a variance of z_ijIs a Gaussian distribution, IG (-) represents an inverse gamma distribution, a₀And b₀Is a hyperparameter of the inverse gamma distribution, which can usually be set simply to 10⁶。

From this, each pixel point x can be obtained_ijProbability density distribution function of

z_ij～IG(a₀，b₀) (1-4)

Generally, a priori information of the whole structure and parts of details can be grasped simultaneously. U and V may be estimated using a Bayesian hierarchical model. In general, U and V can be set to gaussian prior with a mean of zero, and the variance of the gaussian distribution is assumed to follow the gamma distribution, so as to reduce the negative impact of the prior distribution on the estimation result. As shown in FIG. 2, the decomposition model is represented as

τ_u～G(c₀，d₀)，τ_v～G(e₀，f₀)

Wherein, tau_uIs u_iA priori distribution of_vIs v is_jA priori distribution of_uAnd τ_vIs respectively c₀、d₀And f₀For a parametric gamma distribution, it is usually also simple to set it to 10⁶Of course, the initialization values of a0, b0, c0, d0, e0 and f0 can be set to other constants according to requirements.

S103, optimizing the decomposition model by using a variational Bayesian method to obtain a parameter formula corresponding to each parameter of the decomposition model;

considering that the decomposition model is adopted for direct calculation, a large error may exist in precision, and after the decomposition model is obtained, the decomposition model based on the low-rank matrix of the individual Gaussian noise can be optimized by using the variational Bayesian method, so that a parameter formula corresponding to each parameter of the decomposition model is obtained.

S104, iterating all the parameter formulas to obtain values of all parameters in the decomposition model;

in specific implementation, each parameter in the decomposition model may be initialized first, and then all the parameter formulas are iterated to obtain the value of each parameter in the decomposition model.

S105, calculating a vector corresponding to the first matrix and a vector corresponding to the second matrix according to the values of the parameters in the decomposition model to obtain a target image;

in specific implementation, after values of each parameter in the decomposition model are obtained through iteration, a vector corresponding to the first matrix and a vector corresponding to the second matrix can be calculated according to the values of each parameter in the decomposition model, and the target image is obtained, so that noise is removed.

S106, judging whether the quality of the target image meets the standard or not;

after the target image is obtained through reconstruction, the quality of the target image can be calculated, and meanwhile, whether the quality of the target image meets the standard or not can be judged, so that the next operation flow is determined.

If the quality of the target image meets the standard, executing step S107 and outputting the target image;

if the quality of the target image does not meet the standard, step S108 is executed, and iteration is continued until the quality of the target image meets the standard.

In specific implementation, if the quality of the target image meets the standard, it can be considered that the noise in the initial image is removed, and the target image can be directly output as a denoised image. If the quality of the target image does not meet the standard, the noise in the initial image is not completely removed or the removal effect is not ideal enough, and the iteration can be continued until the quality of the target image meets the standard.

The image denoising method based on the heterogeneous noise characteristics provided by this embodiment is implemented by preprocessing image data, establishing a decomposition model that obeys each pixel point in a target data set, and outputting a denoised target image through values of parameters in an iterative decomposition model, so as to improve the calculation efficiency and the denoising effect.

On the basis of the foregoing embodiment, in step S103, optimizing the decomposition model by using a variational bayesian method to obtain a parameter formula corresponding to each parameter of the decomposition model, the method includes:

calculating posterior distribution of each parameter of the decomposition model according to the variational Bayes method;

respectively calculating an iterative formula of each parameter of the decomposition model according to the posterior distribution of each parameter of the decomposition model;

and taking the iterative formula as a parameter formula corresponding to each parameter of the decomposition model.

Further, u in each parameter of the decomposition model_iA posterior distribution of (a) and (v)_jThe posterior distribution of (A) is GaussianThe posterior distribution of the cloth is Gaussian distribution;

τ_ua posterior distribution of (d) and_vthe posterior distribution of (a) is a gamma distribution;

z_ijthe posterior distribution of (a) is an inverse gamma distribution.

In specific implementation, the first matrix u_iPosterior distribution q (u)_i) Is a Gaussian distribution, i.e.

Wherein the content of the first and second substances,is the mean value of the gaussian distribution and,i is an identity matrix of r × r, which is the variance of the gaussian distribution.

Parameter tau_uA posterior distribution q (τ)_u) Is a gamma distribution, i.e.

Wherein the gamma distribution parameterIs the iterative formula of

Low dimensional matrix v_jPosterior distribution q (v)_j) Is a Gaussian distribution, i.e.

WhereinIs the mean value of the gaussian distribution and,i is an identity matrix of r × r, which is the variance of the gaussian distribution.

Parameter tau_vA posterior distribution q (τ)_v) Is a gamma distribution, i.e.

Wherein the parameters of the gamma distributionAndis the iterative formula of

Parameter z_ijPosterior distribution q (z)_ij) Is an inverse gamma distribution, i.e.

Wherein the parameters of the inverse gamma distributionIs the iterative formula of

The decomposition model relates to the expected calculation formula as shown in table 1,

TABLE 1

On the basis of the foregoing embodiment, the determining whether the quality of the target image meets the standard in step S105 includes:

judging whether the root mean square error of the target image is smaller than or equal to a threshold value;

in specific implementation, the threshold may be obtained by measuring a plurality of noiseless sample images, or may be set according to accuracy requirements, after the target image is obtained, the root mean square error between the target image and an expected image may be calculated, and then the root mean square error is compared with the threshold, so as to determine the next operation flow.

If the root mean square error of the target image is smaller than or equal to the threshold, judging that the quality of the target image meets the standard;

and if the root mean square error of the target image is larger than the threshold value, judging that the quality of the target image does not meet the standard.

In specific implementation, if the root mean square error of the target image is less than or equal to the threshold, it may be determined that the target image has achieved an expected effect, and then it is determined that the quality of the target image meets a standard, and if the root mean square error of the target image is greater than the threshold, it may be determined that the target image has not achieved the expected effect, and then it is determined that the quality of the target image does not meet the standard.

The present solution will be described below with reference to a specific embodiment, for example, the decomposition model is VBMFG. The above method can be validated using a medical data set and a face data set at the university of yarrowia. The medical database contained 70 cardiac MRI images of size 190 x 90 and the jerryry face dataset contained 64 different exposure face images of size 192 x 168. To illustrate the effectiveness of model performance, the present disclosure employs 7 comparative models, including CWM, ALM, Dampped Wiberg (DW), PRMF, VBMFL1, VBMFL2, MOG, PMOEP algorithms. The comparison index used in this experiment had the root mean square error and the time required for each iteration. In the experiment, the parameter e of the algorithm provided by the disclosure₀And f₀Are all set to 10⁶. The maximum number of iterations of the model is set to 100.

Firstly, the data set is normalized, and then the normalized data is used as the original input of the model. The method carries out different processing on the two data sets respectively, and for the medical data set, the method sets pixel points which are lost by 30%; for the face data set, the uniform noise of [ -50,50] is added in addition to the pixels with 30% loss set by the method.

Experiments were performed to determine the optimal rank of the medical data set, taking into account that different ranks may be set. As shown in fig. 3, the CWM, ALM, VBFML1, VBFML2, VBMFG, DW algorithms all start to converge with rank 7, so the present disclosure sets the optimal rank for the medical dataset to 7, taking into account that the images in the face dataset at yale university all lie on one four-dimensional subspace. Thus, the present disclosure sets the optimal rank of the face dataset of yale university to 4. The processing effect of different low rank matrix decomposition models on two data sets is shown in table 2, fig. 4 and fig. 5.

TABLE 2

Referring to fig. 6, an embodiment of the present disclosure also provides an electronic device 60, including: at least one processor and a memory communicatively coupled to the at least one processor. Wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method for image denoising based on heterogeneous noise characteristics of the method embodiments described above.

The disclosed embodiments also provide a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the heterogeneous noise characteristic-based image denoising method in the foregoing method embodiments.

The disclosed embodiments also provide a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, cause the computer to perform the heterogeneous noise characteristic-based image denoising method in the aforementioned method embodiments.

Referring now to FIG. 6, a schematic diagram of an electronic device 60 suitable for use in implementing embodiments of the present disclosure is shown. The electronic devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., car navigation terminals), and the like, and fixed terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 6 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.

As shown in fig. 6, the electronic device 60 may include a processing means (e.g., a central processing unit, a graphics processor, etc.) 601 that may perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)602 or a program loaded from a storage means 608 into a Random Access Memory (RAM) 603. In the RAM 603, various programs and data necessary for the operation of the electronic apparatus 60 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.

Generally, the following devices may be connected to the I/O interface 605: input devices 606 including, for example, a touch screen, touch pad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 607 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 608 including, for example, tape, hard disk, etc.; and a communication device 609. The communication means 609 may allow the electronic device 60 to communicate with other devices wirelessly or by wire to exchange data. While the figures illustrate an electronic device 60 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided.

In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication means 609, or may be installed from the storage means 608, or may be installed from the ROM 602. The computer program, when executed by the processing device 601, performs the above-described functions defined in the methods of the embodiments of the present disclosure.

It should be noted that the computer readable medium in the present disclosure can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In contrast, in the present disclosure, a computer readable signal medium may comprise a propagated data signal with computer readable program code embodied therein, either in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.

The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device.

The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to perform the steps associated with the method embodiments.

Alternatively, the computer readable medium carries one or more programs which, when executed by the electronic device, enable the electronic device to perform the steps associated with the method embodiments.

Computer program code for carrying out operations for aspects of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units described in the embodiments of the present disclosure may be implemented by software or hardware.

It should be understood that portions of the present disclosure may be implemented in hardware, software, firmware, or a combination thereof.

The above description is only for the specific embodiments of the present disclosure, but the scope of the present disclosure is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present disclosure should be covered within the scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be subject to the protection scope of the claims.