SYSTEM AND METHOD FOR ASSESSING QUALITY OF DIAGNOSTIC

IMAGES

BACKGROUND

[0001] The subject matter disclosed herein generally relates to image processing. More specifically, the subject matter relates to a method and system for assessing the image quality of diagnostic images, for example medical images.

[0002] Medical imaging for diagnostic purposes have seen a rapid progress with the advancement of techniques including Computer Tomography (CT), Magnetic Resonance Imaging (MRI), nuclear medical imaging such as Single Photon Emission Computed Tomography (SPECT), Positron Emission Tomography (PET) and similar imaging systems. Images generated by such modalities reveal internal anatomy reflecting very complex physical and physiological phenomena. The images can be two-dimensional slices, volume images, temporal stacks or volume over time.

[0003] One of the challenges with such images is that the acquisition of the diagnostic images may introduce degradations. For example, a CT image is degraded by many factors including Poisson noise, whereas the noise in an MRI image is typically characterized as a Rician distribution. The image quality in a CT scanner typically depends on the x-ray tube current (in milli amperes) and exposure time (in seconds). Improved image quality may be achieved in some instances with a proportional increase in the radiation dose to the patient.

[0004] the magnitude of the noise is generally indicated by the variation of pixel intensities over a region of interest (ROI) in a homogeneous substance. Image noise tends to be inversely proportional to the square root of the dose and to the slice thickness. Conventional measures of image quality like signal to noise ratio (SNR), requires an estimate of background noise. However, the background noise estimate may not be readily available and sometimes the estimate may not be accurate. Further, the SNR estimate does not consider the local image characteristics into account.

[0005] There is a need for an enhanced system and method to assess the image quality so that the user can better understand the quality of the image. A superior measure of image quality is desirable in many applications including denoising and other post processing techniques

BRIEF DESCRIPTION

[0006] In accordance with one exemplary embodiment, a method of determining a metric value of a predefined pixel of a diagnostic image is disclosed. The method includes receiving a diagnostic image and identifying a predefined pixel in the diagnostic image. The method further includes computing a plurality of first spatial statistical values of the diagnostic image and determining a homogenous region around the predefined pixel based on the plurality of first spatial statistical values of the diagnostic image. The method further includes computing a plurality of second spatial statistical values of the determined homogeneous region and determining a metric value for the predefined pixel based on the second spatial statistical values. The metric value is representative of a signal to noise ratio of the diagnostic image.

[0007] In accordance with one exemplary embodiment, a system for determining a metric value for a predefined pixel of a diagnostic image is disclosed. The system includes at least one processor-based device having computer instructions to instruct the at least one processor-based device to receive a diagnostic image and to identify a predefined pixel in the diagnostic image. The computer instructions further instruct the at least one processor-based device to compute a plurality of first spatial statistical values of the diagnostic image and to determine a homogenous region around the predefined pixel based on the plurality of first spatial statistical values of the diagnostic image. The computer instructions also instruct the at least one processor-based device to compute a plurality of second spatial statistical values of the determined homogeneous region and to determine a metric value for the predefined pixel based on the second spatial statistical values. The metric value is representative of a signal to noise ratio of the diagnostic image.

[0008] In accordance with another exemplary embodiment, a non-transitory computer readable medium encoded with a program is disclosed. The program instructs at least one processor-based device to receive a diagnostic image and to identify a predefined pixel in the image. The program further instructs at least one processor-based device to compute a plurality of first spatial statistical values of the diagnostic image and to determine a homogenous region around the pixel based on the plurality of first spatial statistical values of the diagnostic image. The program also instructs at least one processor-based device to compute a plurality of second spatial statistical values of the determined homogeneous region and to determine a metric value for the pixel based on the second spatial statistical values.

DRAWINGS

[0009] these and other features and aspects of embodiments of the present invention will become better understood when the following detailed description is read with reference to the accompanying drawings in which like characters represent like parts throughout the drawings, wherein:

[0010] FIG. 1 is a diagrammatic illustration of a medical imaging system in accordance with an exemplary embodiment;

[0011] FIG. 2 illustrates a phantom image in accordance with an exemplary embodiment;

[0012] FIG. 3 illustrates a flow chart for determining a signal quality of a pixel in accordance with an exemplary embodiment;

[0013] FIG. 4 illustrates a grid to be superimposed on a diagnostic image in accordance with an exemplary embodiment;

[0014] FIG. 5 illustrates a cumulative standard deviation (SD) image derived from a diagnostic image in accordance with an exemplary embodiment;

[0015] FIG. 6 illustrates a contour map generated based on a cumulative standard deviation image depicted in the polar coordinate center in accordance with an exemplary embodiment;

[0016] FIG. 7 illustrates a homogenous region extracted from a phantom image in accordance with an exemplary embodiment;

[0017] FIG.8 illustrates a propagated SD image derived from a phantom image in accordance with an exemplary embodiment;

[0018] FIG. 9 illustrates a directional variation analysis map in accordance with an exemplary embodiment;

[0019] FIG. 10A illustrates directional variation analysis map of a noisy image in accordance with an exemplary embodiment; and

[0020] FIG. 10B illustrates a corresponding variation analysis map of the demonized image in accordance with an exemplary embodiment.

DETAILED DESCRIPTION

[0021] Embodiments of the present disclosure relate to a method and system for quantifying the image quality of diagnostic images, for example medical images. The embodiments of the present system may be used to assess the quality of a diagnostic image obtained during acquisition or post processing stages. In one example, a homogenous region around a predefined pixel of the diagnostic image is determined based on a plurality of first spatial statistical values of the diagnostic image. A plurality of second spatial statistical values of the homogenous region is used to determine a metric value representative of signal quality of the predefined pixel, indicative of a quality of the diagnostic image.

[0022] FIG. 1 is an exemplary medical imaging system 100 having a computer tomography (CT) scanner 102 used for performing scanning of a patient 106 positioned on a gantry 104. It should be noted herein that system 100 discussed herein is an exemplary embodiment and should not be construed as limiting the scope of the invention. The CT scanner can be used to scan objects and articles as well as patients that include humans and animals. The configuration of such a system may vary depending on the application. For example, the medical imaging system 100 may also be a Magnetic Resonance (MR) imaging system, a Positron Emission Tomography (PET) imaging system or any other suitable imaging modality.

[0023] In one embodiment, the CT scanner 102 includes a plurality of x-ray sources (not shown) positioned around a bore 116 receiving the gantry 104 and a plurality of detectors (not shown) disposed opposite to the x-ray sources. X-ray beams originating from the sources, are attenuated by the patient 106 and subsequently received by the detectors to generate a diagnostic image 114 of a body portion of the patient 106 to be scanned. A plurality of x-rays are transmitted to different locations of the body portion of the patient 106 through relative motion between the plurality of x-ray sources and the patient position on the gantry 104. The intensity values of image pixels in the diagnostic image 114 are typically proportional to the x-ray beam attenuation represented by CT numbers (measured in Hounsfield Units, HU) of the ingredient materials of the body of the patient 106.

[0024] The image scanner 102 is coupled to a processor-based device 108 which interfaces with the operator, typically through a user interface such as a keyboard 112 and a display 110. The processor-based device 108 may be one or more general purpose processors, controllers, or servers. In one embodiment, the processor-based device 108 may also be embedded wholly or partially in the scanner 102. The processor-based device 108 receives signals output from the scanner 102 and processes the signals to generate the diagnostic image 114 to be displayed on the display 110. The diagnostic image 114 can also be stored or otherwise transmitted to another location for display, further processing or storage. The diagnostic image 114 may be an image acquired by the scanner 102, or a modified image processed by the processor-based device 108. The processor-based device 108 may use software instructions from a disk or from memory. The software may be encoded in any language, including, but not limited to, assembly language, Hardware Description Language, high level languages such as FORTRAN, Pascal, C, C++, python and Java, ALGOL (algorithmic language), and any combination or derivative of at least one of the foregoing languages. The processor-based device 108 may also read instructions from a non-transitory encoded computer medium having instructions to process the diagnostic image 114 in accordance with the exemplary embodiments of the present system.

[0025] The proposed technique of assessing the image quality is performed by the processor-based device 108 in various embodiments outlined in the subsequent paragraphs. The processor-based device 108 may be a noise reduction system or an image post processing module operating on acquired images. In accordance with embodiments of the present technique, a metric value indicative of a quality of the image is determined. The exemplary technique discussed herein provides an assessment of the quality of the image

[0026] The metric value may also be useful in many other scenarios where image processing is involved. In one example, the metric value indicative of a quality of the image is used in an exemplary denoising technique for reducing the effects of noise in the diagnostic image 114. As another example, the proposed technique provides a normalized image quality metric value that can be used to quantify and compare two images. The quality metric value does not need background definition and hence different from the conventional Signal to ratio (SNR) or contrast to noise ratio (CNR). The exemplary technique may also be used to compare the performance of algorithms that affect image quality and to select an optimal method. In an embodiment, reconstruction algorithm used in CT (or in MR or in PET) may be suitably modified to yield optimal performance point based on the determined metric value. A post processing algorithm, employing the exemplary technique, may be able to provide meaningful results in an operating range. The proposed technique may also be used to determine safe operating ranges for attaining image quality from a training set of images. A post processing technique may be used to predict the performance of an algorithm based on the drift in the metric value from the determined safe operating ranges. The metric value may also be used to benchmark a plurality of denoising algorithms. An optimal denoising algorithm may be selected from a plurality of methods by maximizing the metric value based on an iterative optimization technique. Although, specific examples of usage of the metric value are discussed with respect to specific applications, it should be observed herein that the metric value may be used in other ways with equal effectiveness within the scope of the proposed technique.

[0027] FIG. 2 illustrates the phantom image 114 obtained from the system 100 (shown in FIG. 1). Although, the phantom image is illustrated, the present technique is equally applicable to other diagnostic and/or clinical images. It should be noted herein that the terms "phantom image" and "diagnostic image" may be used interchangeably. In one example the diagnostic image 114 has a plurality of regions 202, 204, 206, and 208 representative of a plurality of corresponding ingredient materials of the patient. For example, region 202 is representative of an air region, the region 204 is representative of a calcium region, the region 206 is representative of an iodine region, and the region 208 is representative of a water region. It should be noted herein that generally, a user has no prior knowledge of the number of materials, types of materials, or properties of the materials represented by the diagnostic image 114. The regions 202, 204, 206, and 208 generally have arbitrary shapes and sizes as can be seen in the illustration. But, image pixels within each of these regions 202, 204, 206, and 208 typically exhibit similar visual properties and are generally considered as homogenous regions.

[0028] In accordance with the embodiments of the present system, a Predefined pixel 210 is identified in the diagnostic image 114. It should be noted herein that the predefined pixel 210 may be disposed anywhere in the diagnostic image 114. Further, the predefined pixel 210 may be identified based on at least one of a predefined algorithm or a user input data. Boundary 214 of a homogenous region 212 having the predefined pixel 210 in the image 114 is determined. Thereafter a metric value representative of image quality of the predefined pixel 210 within the homogenous region 212 of the diagnostic image 114 is determined. These determination steps are explained in greater detail herein with reference to subsequent figures.

[0029] FIG. 3 is a flow chart 300 illustrating an exemplary embodiment for determining a metric value indicative of signal quality of the predefined pixel in the diagnostic image. In one example the diagnostic image from the scanner is received by the processor-based device 302. In certain embodiments, the diagnostic image may be generated by a post processing module of the processor-based device adopting a denoising technique. In another embodiment, the diagnostic image may be an intermediate result of an image processing technique. The image can be received from the scanner or from memory representative of a dynamic image or a previously recorded image. A predefined pixel in the diagnostic image is identified in 304 as explained previously with reference to FIG. 2.

[0030] A grid is formed around the predefined pixel in the image 306. The grid has a center point and a plurality of grid lines forming a plurality of bins. A plurality of first spatial statistical values of the diagnostic image is computed in 308 by the processor-based device based on the grid. In one embodiment, the plurality of first spatial statistical values of the diagnostic image may be computed with reference to the predefined pixel without employing the grid. A first spatial statistical value of a bin is determined based on a plurality of image pixels within the bin and is representative of a gradient value across the diagnostic image. For example, the first spatial statistical value of a bin may be based on a mean value of the pixels of the bin. In another example, the first spatial statistical value may be based on a standard deviation (SD) of pixels of the bin. In an exemplary embodiment, a cumulative standard deviation is determined for each of the bins of the grid. An exemplary embodiment of generating the grid and determining cumulative SD values for each of the bins of the grid is explained with reference to FIG. 4.

[0031] A plurality of sectors is formed in the grid with reference to the center point of the grid. Each of the sectors includes a plurality of bins of the grid along a radial direction of the grid. A radial distance value indicative of the contours of a homogenous region around the predefined pixel is identified for each sector based on the cumulative SD values corresponding to the bins of the sector 310. Based on the radial distance values corresponding to each sector, a boundary of the homogenous region is determined 312. Determination of the radial distance values for each sector and determination of the contours of the homogenous region is explained in greater detail with reference to FIG. 6 and FIG. 7.

[0032] A plurality of second spatial statistical values of the diagnostic image is determined for the bins of the determined homogenous region 314. The plurality of second spatial statistical values include values such as mean, variance or standard deviation values corresponding to pixels of the plurality of bins of the homogenous region. In an exemplary embodiment of the present system, the plurality of second spatial statistical values represents directional variation characteristics of the homogenous region. In the presence of non-uniform illumination in the diagnostic image, the directional variation characteristics are determined by computing a propagated standard deviation (SD) value of bins in each of the sectors of the grid as explained in greater detail.

[0033] Based on the propagated standard deviation (SD) values determined for each of the bins of the grid, a directional SD value is determined for each of the sectors of the grid. A signal strength value and a signal perturbation value are determined based on the directional SD values corresponding to the plurality of sectors of the grid 316. The signal strength value is indicative of the directional noise statistics and measures the consistency of a pixel within the neighborhood along each direction. Such a measurement may be utilized to determine the presence of a gaussian noise. A signal perturbation value quantifies the uniformity of noise statistics across all directions. The presence of streak or spike artifact in the homogenous region may be indicated by the signal perturbation value. It should be noted herein that the signal perturbation value exhibits a directional invariance characteristic. A signal confidence value is determined based on the signal strength value and the signal perturbation value of the homogenous region 318. It should be noted herein that the signal confidence value is indicative of the amount of variations in the homogeneous region of the image. The signal confidence value is high when there is no variation within the homogenous region and the signal confidence value is low for textured regions with higher variance. The computation of signal confidence value is based on the foreground pixel values in the best region of interest. Thus, the signal confidence value provides an objective, bounded metric value indicative of a quality of the predefined pixel 210 of the diagnostic image.

[0034] FIG. 4 illustrates an exemplary grid structure 400 used with the diagnostic image for determining the homogenous region around the predefined pixel 210 of FIG. 2. The grid structure 400 includes a plurality of radial lines 402, 404, 406, and 408 passing through a center point 426, and a plurality of circles 418, 420, 422, and 424 with varying radii around the center point 426 of the grid. The plurality of radial lines 402, 404, 406, and 408 and the circular lines 418, 420, 422, and 424 partition the grid 400 into a plurality of sectors 428, 430, 432, 434, 436, 438, 440, and 442. Each sector includes a plurality of bins. For example, the sector 430 has bins 410, 412, 414, and 416. The grid structure 400 is superimposed on the diagnostic image such that the center point 426 of the grid 400 coincides with the predefined pixel 210 of the image 114 of FIG. 2 at which the signal quality is to be determined. When the grid 400 is superimposed on the diagnostic image, each of the bins 410, 412, 414, and 416 of the sector 430 includes a plurality of pixels of the diagnostic image. Similarly, the bins of the other sectors 428, 432, 434, 436, 438, 440, and 442 of the grid 400 include a plurality of pixels of the diagnostic image. It should be noted that the grid illustrated herein is an exemplary embodiment and may vary depending on the imaging application.

[0035] A plurality of first spatial statistical values are computed by determining a cumulative SD value for each of the bins 410,412, 414, and 416 of the grid 400. For the innermost bin 416, a rectangular pixel window 444 around the predefined pixel 210 is considered. The pixel window 444, in one example, may include 3x3 pixels disposed around the predefined pixel 210 disposed coinciding with the center 426 of the grid 400. The size of the pixel window may vary depending on the application. A mean value "u" of pixels in the rectangular window 444 is determined. A cumulative SD (CSD) value of the pixels for the bin 416 is determined by: where Bo represents the bin 416, σ0 is the cumulative SD for the bin B0 with mean value "μ" and xt is a pixel in the bin B0. The cumulative SD for the bin 414 is the SD of the pixels of bins 416 and 414 and represented by: where Bi represents the bin 414, 07 is the cumulative SD for the bin B1 with mean value Μ and x, is a pixel in the bin B0 and B1. Similarly, the cumulative SD values for bins 412 and 410 of the same sector are also determined. Similarly, cumulative SD values of the plurality bins corresponding to all other sectors of the grid 400 are determined.

[0036] FIG. 5 illustrates a cumulative SD (CSD) image 500 corresponding to
the diagnostic image. The value of image pixels in each of the bins of the CSD image 500 corresponds to the CSD value of the respective bins. The cumulative SD image 500 may be represented as a contour plot and boundaries of a homogenous region around the predefined pixel may be determined based on such a contour plot. The identification of the boundaries of the homogenous region around the predefined pixel is explained in greater detail with reference to the subsequent figure.

[0037] FIG. 6 illustrates a contour plot 600 based on the CSD values corresponding to bins of the grid. The x-axis 604 represents radial angle in degrees on the grid from the radial line 408 (shown in FIG. 4) along a clockwise direction. The y-axis 606 represents radial distance on the grid from the predefined pixel 210. The radial distance may be represented by the number of pixels. For a given radial angle, the contours indicate the radial distance along a radial line corresponding to the given radial angle at which substantial gradient differences are observed between successive CSD values. The contour 602 nearest to the x-axis provides the radial distances of the boundaries of the homogenous region around the predefined pixel 210 for all radial angles. The radial distances are determined by processing the CSD values corresponding to the bins of each of the sectors of the diagnostic image. For example, in the sector 430 of FIG. 4, CSD values corresponding to each pair of bins 410 and 412, 412 and 414, and 414 and 416 are compared with each other. A pair of bins having the maximum difference of CSD values is determined. The radial distance from the center point of the grid to the boundary between the pair of bins is represented by the contour 602. Thus, the contour 602 is representative of the boundary of the homogenous region.

[0038] FIG. 7 illustrates a phantom image 700 with the determined homogenous region 702 around the predefined pixel 210. A metric value for the predefined pixel 210 is determined based on the homogeneous region 702 around the pixel 210. A mean value of a plurality of image pixels of the diagnostic image, corresponding to each bin among the plurality of bins, corresponding to each sector, and the center point of the grid is computed. The mean value of the center point is determined to generate a first modified mean value of the first bin. The mean value of the plurality of image pixels of the diagnostic image corresponding to the first bin is determined to generate a second modified mean value of the second bin in each sector. Thus, a plurality of modified mean values is generated. A standard deviation value of the plurality of image pixels corresponding to each bin among the plurality of bins corresponding to each sector, based on the modified mean value of the corresponding bin is determined to generate a plurality of standard deviation values corresponding to each sector.

[0039] As an example, again referring to FIG. 4, for the bin 416, a standard deviation of the pixels of the bin 416 is determined based on a mean value computed from the pixels around the center point 426 which is superimposed on the predefined pixel 210. For the next bin 414, the propagated SD is determined by: ^1= E(**i(0-Aao)2 (3) where, B] represents the bin 414, o\ is the propagated standard deviation corresponding to the bin Bi, xBi(i) is the ith pixel of the bin Bi, fxBo is the mean of the bin BO which is radially inward to the bin Bj.

[0040] FIG. 8 illustrates a propagated SD image 800 derived from the image 114 of FIG. 2 in accordance with an embodiment. Each pixel of the image 800 is a propagated SD value of the plurality of bins of the grid superimposed on the diagnostic image. A median value corresponding to each sector is generated based on the plurality of standard deviation values corresponding to each sector, to generate a plurality of median values corresponding to the plurality of sectors. As an example, with reference to FIG. 4, a directional SD for the sector 430 having the bins 410, 412, 414, and 416 is determined based on the propagated SD values corresponding to the bins 410, 412, 414, and 416. Specifically, the directional SD value is represented by: where, σθ is the directional SD value of a sector θ, σθ(0), σθ(l),.., σθ(k) are propagated SD values of bins of the sector 6, and k-1 is the number of bins in the sector θ in the grid. The directional SD value determined by Equation 4 is not normalized due to variation of intensities across different regions. A normalized directional SD value may be determined as: where σNθ is the normalized directional SD value, σθ is the un-normalized directional SD value, and Sg is a normalizing constant equal to the signal dynamic range within the sector.

[0041] FIG. 9 illustrates a directional variation analysis map 900 generated based on normalized directional values for the homogenous region 702 (shown in FIG. 7). The homogenous region 702 has a plurality of sectors 902 with corresponding direction SD values represented by different patterns.

[0042] A signal strength value based on the plurality of median values corresponding to the plurality of sectors is determined. Similarly, a signal perturbation value based on the plurality of median values corresponding to the plurality of sectors is also determined. The signal strength value Ss for the pixel positioned in the homogenous region is represented by: where, σNθ, σNθ2,.....σNθmm are the normalized directional values corresponding to the sectors 61, 62,,..., dm respectively, and Mean is the mean operator. Similarly, the signal perturbation value Sp for the pixel positioned in the homogenous region is represented by: where, σNθ, σNθ ...σNθm are the normalized directional values corresponding to the sectors θ, θ,..., θm respectively, and MAD is the Mean Absolute Deviation operator.

[0043] The metric value representing the signal quality in this example is determined based on the signal strength value Ss determined by Equation 6, and the signal perturbation value Sp determined by the Equation 7. The metric value is determined based on the signal strength value and the signal perturbation value for the predefined pixel in the homogenous region and represented by: where Sc is the metric value (may also be referred to as a signal confidence value), Ss is the signal strength value, and Sp is the signal perturbation value corresponding to the image w.r.t to the pixel in the homogenous region. The signal strength value (Ss) indicates the average variations observed in the homogeneous region. The signal strength value (Ss) value is bounded between [0,1] due to normalization factor Sg of Equation 5. The maximum value of "one" indicates that there is no variation in the homogeneous region of the image. With increase in noise, the variation in the homogenous region increases and the value of the signal strength value decreases towards "zero". Similarly, the signal perturbation value (Sp) is also bounded between [0, 1]. Thus, the signal confidence value, which is the product of the signal strength value and the signal perturbation value, is also bounded between [0, 1]. The signal confidence value of a superior quality image is near "one" and noisier image has a signal confidence value near to "zero".

[0044] FIG. 10A illustrates directional analysis map 1002 of a noisy image 1000 for a CT phantom image. FIG. 10B illustrates corresponding denoised image 1004 in accordance with an exemplary embodiment. Edges 1008 of the denoised image 1004 of FIG. 10B is relatively smooth in comparison with edges 1006 of the noisy image 1000 of FIG. 10A. A signal quality metric calculated for a given voxel is represented as a numerical value at bottom right of each figure. The signal confidence value for the denoised image 1004 is 0.955 which is higher than the confidence value of 0.931 for the noisy image 1000 of FIG. 10A.

[0045] In accordance with the embodiments discussed herein, the signal confidence value is representative of a signal quality of the predefined pixel in a diagnostic image. The signal confidence value is indicative of a normalized image quality with reference to a predefined pixel of the diagnostic image and is suitable for comparing images generated from multiple scanners or to evaluate a plurality of image enhancement methods.

[0046] It is to be understood that not necessarily all such objects or advantages described above may be achieved in accordance with any particular embodiment. Thus, for example, those skilled in the art will recognize that the systems and techniques described herein may be embodied or carried out in a manner that achieves or optimizes one advantage or group of advantages as taught herein without necessarily achieving other objects or advantages as may be taught or suggested herein.

[0047] While the invention has been described in detail in connection with only a limited number of embodiments, it should be readily understood that the invention is not limited to such disclosed embodiments. Rather, the invention can be modified to incorporate any number of variations, alterations, substitutions or equivalent arrangements not heretofore described, but which are commensurate with the spirit and scope of the invention. Additionally, while various embodiments of the invention have been described, it is to be understood that aspects of the invention may include only some of the described embodiments. Accordingly, the invention is not to be seen as limited by the foregoing description, but is only limited by the scope of the appended claims. What is claimed as new and desired to be protected by Letters Patent of the United States is:

CLAIMS:

1. A method, comprising: receiving a diagnostic image; identifying a predefined pixel in the diagnostic image; computing a plurality of first spatial statistical values of the diagnostic image; determining a homogenous region around the predefined pixel based on the plurality of first spatial statistical values of the diagnostic image; computing a plurality of second spatial statistical values of the determined homogeneous region; and determining a metric value for the predefined pixel based on the second spatial statistical values; wherein the metric value is representative of a signal to noise ratio of the diagnostic image.

2. The method of claim 1, further comprising generating a grid having a plurality of bins, and superimposing the grid on the diagnostic image such that a center point of the grid coincides with the predefined pixel of the diagnostic image.

3. The method of claim 2, wherein computing the plurality of first spatial statistical values comprises computing at least one of a mean, a variance, and a standard deviation of a plurality of image pixels of the diagnostic image, corresponding to the plurality of bins.

4. The method of claim 2, wherein computing the plurality of second spatial statistical values comprises computing at least one of a mean, a variance, and a standard deviation of a plurality of image pixels of the diagnostic image, corresponding to the plurality of bins.

5. The method of claim 2, wherein generating the grid comprises having a plurality of concentric circles and a plurality of radial lines extending across the plurality of concentric circles to form a plurality of sectors, wherein each sector comprises the plurality of bins arranged along a radial direction of the grid.

6. The method of claim 5, wherein determining the homogenous region comprises: determining a gradient value for each bin of the grid to generate a plurality of gradient values; determining a radial distance value of each sector based on the plurality of gradient values for the plurality of bins corresponding to each sector to generate a plurality of radial distance values; and determining the homogenous region around the predefined pixel based on the plurality of radial distance values.

7. The method of claim 5, comprising forming the plurality of bins comprising at least a first bin and a second bin disposed radially outward with reference to the first bin in each sector.

8. The method of claim 7, wherein computing the second spatial statistical values comprises: computing a mean value of a plurality of image pixels of the diagnostic image, corresponding to each bin among the plurality of bins, and a center point of the grid, corresponding to each sector, to generate a plurality of mean values corresponding to each sector; determining a first modified mean value of the plurality of image pixels of the diagnostic image, corresponding to the first bin, based on the corresponding mean value of the center point, and a second modified mean value of the plurality of image pixels of the diagnostic image, corresponding to the second bin, based on the corresponding mean value of the first bin, to generate a plurality of modified mean values; and determining a standard deviation value of the plurality of image pixels corresponding to each bin among the plurality of bins corresponding to each sector, based on the modified mean value of the corresponding bin, to generate a plurality of standard deviation values corresponding to each sector.

9. The method of claim 8, wherein determining the metric value comprises: computing a median value corresponding to each sector based on the plurality of standard deviation values corresponding each sector, to generate a plurality of median values corresponding to the plurality of sectors; determining a signal strength value based on the plurality of median values corresponding to the plurality of sectors; determining a signal perturbation value based on the plurality of median values corresponding to the plurality of sectors; and determining the metric value based on the signal strength value and the signal perturbation value.

10. The method of claim 1, wherein identifying the predefined pixel comprises identifying an image pixel based on at least one of a predefined algorithm or a user input data.

11. A system, comprising: at least one processor-based device having computer instructions, wherein the computer instructions instruct the at least one processor-based device to: receive a diagnostic image; identify a predefined pixel in the diagnostic image; compute a plurality of first spatial statistical values of the diagnostic image; determine a homogenous region around the predefined pixel based on the plurality of first spatial statistical values of the diagnostic image; compute a plurality of second spatial statistical values of the determined homogeneous region; and determine a metric value for the predefined pixel based on the second spatial statistical values; wherein the metric value is representative of a signal to noise ratio of the diagnostic image.

12. The system of claim 11, wherein the computer instructions further instruct the at least one processor-based device to generate a grid having a plurality of bins, and superimpose the grid on the diagnostic image such that a center point of the grid coincides with the predefined pixel of the diagnostic image.

13. The system of claim 12, wherein the computer instructions further instruct the at least one processor-based device to compute the plurality of first spatial statistical values comprising at least one of a mean, a variance, and a standard deviation of a plurality of image pixels of the diagnostic image, corresponding to the plurality of bins.

14. The system of claim 12, wherein the computer instructions further instruct the at least one processor-based device to compute the plurality of second spatial statistical values comprising at least one of a mean, a variance, and a standard deviation of a plurality of image pixels of the diagnostic image, corresponding to the plurality of bins.

15. The system of claim 12, wherein the computer instructions instruct the at least one processor-based device to generate the grid having a plurality of concentric circles and a plurality of radial lines extending across the plurality of concentric circles to form a plurality of sectors, wherein each sector comprises the plurality of bins arranged along a radial direction of the grid.

16. The system of claim 15, wherein the computer instructions instruct the at least one processor-based device to determine a gradient value for each bin of the grid to generate a plurality of gradient values; determine a radial distance value of each sector based on the plurality of gradient values for the plurality of bins corresponding to each sector to generate a plurality of radial distance values; and determine the homogenous region around the predefined pixel based on the plurality of radial distance values.

17. The system of claim 15, wherein the computer instructions instruct the at least one processor-based device to form the plurality of bins comprising at least a first bin and a second bin disposed radially outward with reference to the first bin in each sector.

18. The system of claim 17, wherein the computer instructions instruct the at least one processor-based device to: compute a mean value of a plurality of image pixels of the diagnostic image, corresponding to each bin among the plurality of bins, and a center point of the grid, corresponding to each sector, to generate a plurality of mean values corresponding to each sector; determine a first modified mean value of the plurality of image pixels of the diagnostic image, corresponding to the first bin, based on the corresponding mean value of the center point, and a second modified mean value of the plurality of image pixels of the diagnostic image, corresponding to the second bin, based on the corresponding mean value of the first bin, to generate a plurality of modified mean values; and determine a standard deviation value of the plurality of image pixels corresponding to each bin among the plurality of bins corresponding to each sector, based on the modified mean value of the corresponding bin, to generate a plurality of standard deviation values corresponding to each sector.

19. The system of claim 18, wherein the computer instructions instruct the at least one processor-based device to: compute a median value corresponding to each sector based on the plurality of standard deviation values corresponding each sector, to generate a plurality of median values corresponding to the plurality of sectors; determine a signal strength value based on the plurality of median values corresponding to the plurality of sectors; determine a signal perturbation value based on the plurality of median values corresponding to the plurality of sectors; and determine the metric value based on the signal strength value and the signal perturbation value.

20. A non-transitory computer readable medium encoded with a program to instruct at least one processor-based device to: receive a diagnostic image; identify a predefined pixel in the image; compute a plurality of first spatial statistical values of the diagnostic image; determine a homogenous region around the pixel based on the plurality of first spatial statistical values of the diagnostic image; compute a plurality of second spatial statistical values of the determined homogeneous region; and determine a metric value for the pixel based on the second spatial statistical values.