News: Imatest 23.1 (March 2023) (available in the Imatest Pilot program). New methods for calculating camera information capacity, Noise Power Spectrum (NPS), Noise Equivalent Quanta (NEQ), and Ideal observer SNR (SNRi) from slantededge patterns are now available. 
The basic premise of this work is that Information capacity is a superior Shannon information capacity can also be calculated from images of the Siemens star, introduced in 2020.
The Siemens star method was presented at the Electronic Imaging 2020 conference, and published in the paper, “Measuring camera Shannon information capacity from a Siemens star image” and announced in the Imatest News Post: Measuring camera Shannon information capacity with a Siemens star image. The 2020 white paper, Camera information capacity: a key performance indicator for Machine Vision and Artificial Intelligence systems, is a more readable introduction to the Siemens Star measurement. This page presents new calculations of information capacity and additional measurements from slanted edges. 
Introduction – Edge variance calculation – Noise image calculation – Acquiring and framing – Running the MTF module
Results – Edge/MTF plot – Edge/noise plot – 3D plot and Total information capacity – Noise image results –
Noise Power Spectrum (NPS) – Noise Equivalent Quanta (NEQ) – Noise Autocorrelation – SNRi
Original & Noise image – Summary – Links
Calculation summary page: Introduction – Edge variance calculation – Noise image calculation
Meaning of Information capacity
This page contains instructions for Imatest’s new slantededge
information capacityrelated measurements.
The equations, algorithms, and meaning of information capacity are on a separate page.
Introduction to the new measurements
This page contains instructions for obtaining image quality measurements introduced in Imatest 23.1, released in March 2023. These measurements take advantage of newlydiscovered properties of slanted edges– Imatest’s most widely used patterns for measuring MTF. Most are unfamiliar because they have been siloed in the medical imaging community, and have beentraditionally difficult to perform. We describe two new measurement methodologies.
 The Edge variance method, conveniently calculates camera information capacity.
 The Noise image method uses a different approach to calculate information capacity and several additional image quality factors.
This page introduces the new calculations, shows how to obtain them, then presents Key Results.
Detailed algorithms are presented in Information capacity measurements from Slanted edges: Equations and Algorithms.
.
A: Edge variance calculation summary
Sum the squares of the scan lines to obtain the edge variance.
The Edge variance Information capacity calculation is described in detail in the white paper, Measuring Camera Information Capacity with slanted edges: the Edge variance calculation and in the paper presented at Electronic Imaging 2023, “Measuring Camera Information Capacity with Slanted Edges” (now on the Imatest website: we’ll use the IS&T EI2023 link when it’s available). The descriptions on this page are very concise; there is more detail in the two linked documents.
Starting with images of slanted edges, (Original ROI, below), typically made from a 4:1 contrast chart (chart contrasts between 2:1 and 10:1 are acceptable), the standard slantededge algorithm
 finds the center of each scan line (taken in a direction approximately normal to the edge),
 fits the centers to a polynomial curve, then,
 depending on the location of the polynomial at the scan line, it adds the contents of the scan line to one of four bins.
 These mean values of each bin are then interleaved to form a 4× oversampled edge, which has lower noise than any of the individual scan lines.
The Edge variance method adds another summation. The squares of each scan line are added to the appropriate bin to calculate the variance of the edge signal, σ^{2}, which is equivalent to its noise power, N. The mean signal amplitude for a uniformlydistributed signal of peaktopeak amplitude V_{PP} is \(S(f) = V_{PP} MTF(f) / \sqrt{12}\) — a reasonable number to use for the information capacity calculation using the ShannonHartley equation, shown above.
The key results are,
C_{4} is the Shannon capacity measured directly from (usually) 4:1 contrast ratio slantededges. It is calculated from the Shannon Hartley equation, using several assumptions (that the signal is uniformly distributed over the peaktopeak signal measurement and the noise power spectral density (NPD) is flat). C_{4} is a special case of C_{n}, for a n:1 contrast ratio (with ISO standard 4:1 strongly recommended) C_{n} is sensitive to chart contrast ratio and exposure, making it interesting for measuring performance as a function of exposure but not robust for calculating a camera’s maximum information capacity.
C_{max} is a much more stable measurement of the maximum information capacity for the camera starting with C_{4} (for the 4:1 contrast chart). It is also insensitive to exposure, at least for linear sensors, where noise is a known function of signal voltage.
B. Noise image calculation summary
Subtract a lownoise debinned / reverseprojected ROI image to obtain a noise image.
This method involves inverting the ISO 12233 binning procedure. Noting that the 4× oversampled edge was created by interleaving the contents of 4 bins, we apply an inverse of the binning algorithm to set the contents of each scan line to its corresponding bin (Inverse binned… ROI, below). Since the inversebinned image is nearly noiseless, a noise image can be created by subtracting the inversebinned image from the original image. This image, adjusted to make the mean (zero) value middle gray, is illustrated as the Noise image ROI, below.
Measurement  Description 
Noise Power Spectrum, NPS(f)  NPS was implicitly assumed to be constant (white noise) in the Edge variance method. 
Noise Equivalent Quanta, NEQ(f) and NEQ_{info}(f) 
measures of frequencydependent signaltonoise ratio (SNR). \(NEQ(f) = \mu^2\ MTF^2(f) / NPS(f)\text{, where }\mu = V_{mean}\) has been used for quantifying medical image quality, but are much less familiar in general imaging. NEQ(f) is equivalent to the number of quanta detected by the sensor when photon shot noise is dominant. It is appropriate for calculating Digital Quantum Efficiency (DQE), when the density of quanta reaching the image sensor is known. NEQ_{info}(f) is derived from \(\mu = V_{PP} / \sqrt{12}\), making it wellsuited for calculating information capacity C_{NEQ}. 
Information capacities C_{4NEQ} and C_{maxNEQ} 
correspond to C_{4} and C_{max} from the Edge variance method, but are derived from NEQ_{info}(f). They are close, but not identical. 
Ideal observer SignaltoNoise Ratio, SNRi 
From Skorka and Kane [9], “The Ideal Observer is a Bayesian decision maker that maximizes the statistical precision of a hypothesis test with two possible outcomes.” SNRi as we present is, is a metric of the detectability of small objects (squares or rectangles), typically of low contrast. 
Noise autocorrelation  The inverse Fourier transform of the Noise Voltage Spectrum. Related to sensor electrical crosstalk. 
Why two methods?The Edge variance was developed first, starting in late October 2022. It was presented at the Electronic Imaging 2022 conference. The Noise image was developed in February 2023, about a month after the conference. It measures more image quality parameters than the Edge variance method. We currently recommend the Noise image method because we believe the camera information capacity measurement is slightly more accurate. The reason: it calculates the Noise Power Spectrum, while the Edge variance assumes the NPS is flat (white). On the other hand, The Edge variance method can provide useful (if imperfect) results for bilateralfiltered images (most JPEGs from consumer cameras), while the Noise image method is only recommended for uniformly or minimallyprocessed images. 
Acquire the image
Acquire a wellexposed image of any slantededge test chart (SFRplus, eSFR ISO, Checkerboard, SFRreg, or SFR) in even, glarefree light. Chart edge contrast ratio should be 4:1 (the ISO 12233 standard) where possible. The limits are 2:1 to 10:1. For the most part, acquiring the image is identical to standard MTF measurements.
Exposure consistency is important because exposure affects the information capacity measurement. (Exposure has only a secondorder effect on MTF measurements.) The mean pixel level of a perfectly exposed slantededge region should be in the range of 0.20 to 0.26, but there is plenty of latitude as long is the image isn’t saturating.
For best results the slanted edge should be 100 pixels or more along the edge. Fewer pixels will work, but the results will be less consistent.
Run the MTF calculation module
Run either an appropriate Rescharts module (interactive; recommended for starting and making settings — SFRplus Setup, eSFR ISO Setup, SFRreg Setup, or Checkerboard Setup in the second column in the Imatest main window or an appropriate fixed, batchcapable module (SFR, SFRplus Auto, eSFR ISO Auto, SFRreg Auto, or Checkerboard Auto buttons in the left column).
Edge/MTF display — In the Settings window, select the appropriate setting in the Information capacity display dropdown menu. It may be somewhat inconspicuous. Calculating information capacity slows down operations slightly. Two examples are shown below.
information capacity noise calculation setting, from a crop of the Rescharts Settings window 
Information capacity noise calculation setting, from left side of More settings window 
The full windows and complete instructions are in SFRplus, eSFR ISO, Checkerboard, SFRreg, or SFR.
The reason for the different Edge variance noise calculations is
 uniformly or minimallyprocessed images, often converted in the computer from raw images, pictured here, have relatively uniform noise, which is somewhat signaldependent. There is no noise peak near the transition. Hence the mean edge noise in the ROI is most appropriate.
 bilateralfiltered images, including most JPEGs from consumer cameras, pictured here, are sharpened near the transition and noisereduced (lowpassfiltered) elsewhere. Hence the average noise would be too low to represent system performance. Hence the (smoothed) peak noise is most appropriate.
 The presence or absence of a peak can be used to detect bilateral filtering (for setting 2: Auto edge noise detection), but it’s not entirely accurate since soft images may exhibit little peak.
If an information capacity calculation is selected (25, below), both Edge variance and Noise image calculations are performed. These settings affect the Edge/MTF display, which has limited space. Results from both methods are displayed in the Noise Spectrum, NEQ, SNRi plot when Results summary is selected for the lower plot.
#  Setting  Description  Best for 
1.  No information capacity calculation  slightly faster than calculating information capacity. About the only time it’s appropriate is for highspeed realtime image acquisition.  Fastest, but not by much since info cap calculations are turned off during reloads. 
2.  Calculate info cap: Auto edge noise detection  Selects the noise calculation method depending on the detection (presence or absence) of a noise peak. Often recommended, but we have found that it sometimes misses the peak in bilateralfiltered images. One of the next two choices is recommended if you predominantly use one file type (either minimally/uniformly processed or bilateralfiltered).  for unknown image processing: “black box” 
3.  Calculate info cap using mean edge noise (default) (Method (1))  for images converted from raw with minimal or uniform processing (no bilateral filtering).  for uniformly / minimallyprocessed images 
4.  Calculate info cap using smoothed peak noise (Method (2))  for bilateralfiltered images, which includes most JPEG images from consumer cameras. It works for uniformlyprocessed images, but gives less stable results than Method (1).  for bilateralfiltered images (most consumer camera JPEGs) 
5.  Calculate info cap – Auto & NEQ Display NEQ  Display the results of the Noise image method in the Edge/MTF plot. Display the results of both methods, with Auto edge noise detection for the Edge variance plots, in the Edge/MTF display. This is our current recommendation for minimally/uniformly processed images.  Best all around 
When OK is pressed the image is analyzed. Any of several displays can be selected in Rescharts, but only a few (shown below) are related to information capacity. The same results can be obtained from running fixed modules.
From the Rescharts interactive interface, the noise calculation can always be set (or changed) from a dropdown menu on the left of the More settings window. This setting is also in the SFR and Rescharts SFR Settings windows.
In the lowerleft of the More settings window, choose one of the Imatest edge calculations, preferably Imatest 22.1 (recommended). At this writing it produces the most reliable results.
Key results
Edge/MTF plot
Information capacity (for the individual edge) has been added to the upper (Edge) plot. Nothing else has changed.
The image below is from a rawconverted eSFR ISO chart image from a high quality compact camera.
Edge/MTF Results from eSFR ISO image converted from raw with minimal processing
Both information capacity C_{4} (for 4:1 contrast charts; very sensitive to exposure) and C_{max} (maximum information capacity; relatively insensitive to exposure) are displayed.
Edge & Info Capacity Noise
This is a variant of the Edge and MTF plot. The top plot is similar (though the Line Spread Function — the derivative of the edge — is of special interest for this plot).
Line Spread Function (LSF) and signaldependent noise σ from
eSFR ISO image converted from raw with minimal processing
The solid line is the noise smoothed by a rectangular kernel with width = 5 (4× oversampled) pixels (1.25 original pixels). Note that the noise is very rough and that there is no distinct peak near the edge transition. From observing regions of the chart, the bumps on the noise curve appear to be random and not repeatable.
Now it gets interesting. The image used for the above plots is derived from a raw image with minimal processing (straight gamma curve; no sharpening or noise reduction — and definitely NO bilateral filtering). Here are results for the incamera JPEG from the same acquisition. Note the large noise peak near the edge. We use Method (1) to calculate noise for the ShannonHartley equation, which takes the average over a fairly wide region near the edge.
Line Spread Function (LSF) and signaldependent noise σ from
eSFR ISO image captured as a JPEG (with strong sharpening and bilateral filtering)
Method (2): Calculate noise for info cap using the maximum value of the smoothed edge noise. This method is recommended for bilateralfiltered JPEGs because it uses a relatively narrow region near the noise peak for measuring noise. Information capacity C_{4} = 3.11 b/p is only slightly higher than for the TIFF: 3.0 b/p. If Method (1) (averaging over a large area, which doesn’t emphasize the peak) is chosen, C_{4} = 3.84 b/p: significantly larger than the raw measurement, and clearly not accurate.
3D plot and total information capacity
C_{4} and C_{max} have been added to the available 3D plots.
3D plot of Edge Information capacity C_{4}
The total information capacity C_{total} (referenced to the signal level: in this case 4:1) is calculated by multiplying the weighted mean (with all weights are set to 1 for the two information capacity plots) in bits per pixel by the total number of megapixels. It is displayed next to the mean information capacity (in bits per pixel).
In the above image, C_{total} = 2.847 bits/pixel × 16 megapixels = 45.44 megabits. The display shows Edge Info Cap C_{max} Mean: 2.847; total Mb: 45.44.
Results for Noise image calculation: NPS, NEQ, and SNRi
The new results (NPS, NEQ, SNRi) and an old one (MTF) can be displayed in the 9. Noise Spectrum, NEQ, SNRi plot. This plot shows results for all available color channels. A 24 megapixel Micro FourThirds camera with a high quality 60mm lens set to f/8 and ISO 100 was used to obtain these results.
Noise Power Spectrum (NPS) plot and results summary from Rescharts
There are several independent choices for the upper and lower plots in two dropdown menus on the right of the figure.
Upper plot  Lower plot 
Noise Voltage Spectrum *for checking inputs to NEQ calculation 
Original image crop 
The checkbox between the two dropdown menus (set to Log xaxis above) lets you select a linear or logarithmic xaxis in the upper plot.
Upper plots
For plots with frequency in the xaxis, the Frequency display can be set to linear or logarithmic (Log xaxis is checked in the Display area, above).
Noise Power or voltage Spectrum (NPS)The NPS (upper) plot is shown above with a logarithmic xaxis and on the right with a linear xaxis (selected by a checkbox). The Noise Power and Voltage Spectrum plots have the same shape: only the yaxis labels are different. The 1D Noise Power or Voltage spectrum is derived from a 2D Fourier transform (FFT) of the noise image. The initial 2D FFT has zero frequency at the image center. The image is divided into several annular regions, and the average noise power is found for each region. NPS is primarily used to check the input to the NEQ plot. 

Noise Equivalent Quanta (NEQ)The standard NEQ used in the plot is based on the mean signal voltage, V_{mean}, shown above. A different voltage, VPP /sqrt(12), is used to calculate NEQbased information capacity, C_{NEQ}. The NEQ plot is rough because of the relatively small size of the slantededge ROIs (Regions of interest). 

Modulation Transfer Function (MTF)Available here for convenience. It looks a little different from the standard MTF plot because the yaxis is logarithmic. 

Edge voltage unnormalizedDisplayed primarily to check input for the NEQ calculation. It’s often interesting to compare peaktopeak amplitudes of the different channels. 

Noise autocorrelation (related to sensor electrical crosstalk)This plot has been included more for R&D than for regular testing. The author (NLK) added it to examine his hypothesis that the noise power spectrum (and autocorrelation) indicate the amount of electrical crosstalk of image sensor when the effects of demosaicing and fixedpattern noise are removed (not the case for the image on the right) and the primary noise source is photon shot noise. The idea behind the hypothesis is that light incident on the sensor is entirely uncorrelated, so that if there were no crosstalk the noise would be white. This image on the right was whitebalanced. The curve is the inverse Fourier transform of the noise spectrum, based on the author’s limited understanding of the WienerKhinchin theorem. 

The image on the right is not WhiteBalanced. The red channel has a larger autocorrelation distance than the other channels, as we would expect. Click on the image to enlarge it. A similar autocorrelation plot can also be obtained from a flat field image in the Image Statics module. We were concerned abuit a relatively large autocorrelation at large distances until we corrected the illumination nonuniformity. 
Lower plots
Ideal observer SignaltoNoise Ratio (SNRi)SNRi is an indicator of the detectability of small objects. It can be displayed for three feature types: a square of sides w and rectangles of sizes w × 2w and w × 4w for w = the smaller side in pixels. SNRi is a new addition to Imatest. We’ll have more to say about it when we gain experience in using and interpreting it. It is discussed in detail in papers by Paul Kane [8] and Orit Skorka and Paul Kane [9]. 

Image crops (original, unbinned, noise)These full image crops are shown below. We illustrate crops here for a noisy image (ISO 12800), where the difference between the original and debinned image is easy to see. The images are all gammaencoded for display. The noise image on the right has been lightened and contrastboosted. The effectiveness of the debinning for removing noise is impressive.



Results summaryThe first two lines contain a summary of image properties. The bottom group of lines contain several variables used in program development to check the accuracy of the calculations. The primary results are the Information capacities in the middle. C_{4}(EdgeVar) and C_{max}(EdgeVar) are derived from the Edge Variance method. C_{4}(NEQ) and C_{max}(NEQ) are derived from the the Noise image method (designated NEQ because they are calculated directly from NEQ). They are close but not identical because the Noise image method uses the measured Noise Power Spectrum, while the Edge Variance method assumes white noise power. 

Feature (Square) visibility 
Summary
 Shannon information capacity C has long been used as a measure of the goodness of electronic communication channels. It specifies the maximum rate at which data can be transmitted without error if an appropriate code is used (it took nearly a halfcentury to find codes that approached the Shannon capacity). Coding is not an issue with imaging. Rodney Shaw’s paper from 1962 [10] is a particularly good example of measuring C for photographic film— it wasn’t easy back then.
 C is ordinarily measured in bits per pixel. The total capacity is \( C_{total} = C \times \text{number of pixels}\).
 The channel must be linearized before C is calculated, i.e., an appropriate gamma correction (signal = pixel level gamma, where gamma ~= 2 for images in standard color spaces such as sRGB or Adobe RGB) must be applied to obtain correct values of S and N. The value of gamma (close to 2) can be determined from runs of any of the Imatest modules that analyze grayscale step charts: Stepchart, Colorcheck., Color/Tone, Multitest, SFRplus, or eSFR ISO. But in most cases it can be determined from the edge image if the chart contrast is entered and Use for MTF is checked.
 We hypothesize that C can be used as a figure of merit for evaluating camera quality, especially for machine vision and Artificial Intelligence cameras. (It doesn’t directly translate to consumer camera appearance because they have to be carefully tuned to reach their potential, i.e., to make pleasing images). It provides a fair basis for comparing cameras, especially when used with images converted from raw with minimal processing.
 Imatest calculates the Shannon capacity C for the Y (luminance; 0.212*R + 0.716*G + 0.072*B) channel of digital images, which approximates the eye’s sensitivity. It also calculates C for the individual R, G, and B channels as well as the C_{b} and C_{r} chroma channels (from YC_{b}C_{r}).
 Shannon capacity has not been used to characterize photographic images because it was difficult to calculate and interpret. But now it can be calculated easily, its relationship to photographic image quality is open for study.
 Since C is a new measurement, we are interested in working with companies or academic institutions who can verify its suitability for Artificial Intelligence systems.
Note: A slantededge information capacity measurement used prior to Imatest 2020, used primarily to obtain total information capacity from Siemens star measurements, has been deprecated completely because it was not sufficiently accurate. 
Links (more links in the White Paper)
 C. E. Shannon, “A mathematical theory of communication,” Bell Syst. Tech. J., vol. 27, pp. 379–423, July 1948; vol. 27, pp.
623–656, Oct. 1948.  C. E. Shannon, “Communication in the Presence of Noise”, Proceedings of the I.R.E., January 1949, pp. 1021.
 Wikipedia – Shannon Hartley theorem has a frequency dependent integral form of Shannon’s equation that is applied to both Imatest’s sine pattern and slanted edge Shannon information capacity calculation.
 I.A. Cunningham and R. Shaw, “Signaltonoise optimization of medical imaging systems”, Vol. 16, No. 3/March 1999/pp 621632/J. Opt. Soc. Am. A
 Brian W. Keelan, “Imaging Applications of Noise Equivalent Quanta” in Proc. IS&T Int’l. Symp. on Electronic Imaging: Image Quality and System Performance XIII, 2016, https://doi.org/10.2352/ISSN.24701173.2016.13.IQSP213.
 Michail C, Karpetas G, Kalyvas N, Valais I, Kandarakis I, Agavanakis K, Panayiotakis G, Fountos G., Information Capacity of Positron Emission Tomography Scanners. Crystals. 2018; 8(12):459. https://doi.org/10.3390/cryst8120459.
 Christos M. Michail, Nektarios E. Kalyvas, Ioannis G. Valais, Ioannis P. Fudos, George P. Fountos, Nikos Dimitropoulos, Grigorios Koulouras, Dionisis Kandris, Maria Samarakou, Ioannis S. Kandarakis, “Figure of Image Quality and Information Capacity in Digital Mammography”, BioMed Research International, vol. 2014, Article ID 634856, 11 pages, 2014. https://doi.org/10.1155/2014/634856.
 Paul J. Kane, “Signal detection theory and automotive imaging”, Proc. IS&T Int’l. Symp. on Electronic Imaging: Autonomous Vehicles and Machines Conference, 2019, pp 271 – 278, https://doi.org/10.2352/ISSN.24701173.2019.15.AVM027.
 Orit Skorka, Paul J. Kane, “Object Detection Using an Ideal Observer Model”, IS&T Int’l. Symp. on Electronic Imaging: Autonomous Vehicles and Machines, 2020, pp 411 – 417, https://doi.org/10.2352/ISSN.24701173.2020.16.AVM041.
 R. Shaw, “The Application of Fourier Techniques and Information Theory to the Assessment of Photographic Image Quality”, Photographic Science and Engineering, Vol. 6, No. 5, Sept.Oct. 1962, pp.281286. Reprinted in “Selected Readings in Image Evaluation,” edited by Rodney Shaw, SPSE (now SPIE), 1976. A fascinating and difficult calculation of information capacity of photographic film. Available for download
 X. Tang, Y. Yang, S. Tang, Characterization of imaging performance in differential phase contrast CT compared with the conventional CT: Spectrum of noise equivalent quanta NEQ(k), Med Phys. 2012 Jul; 39(7): 4467–4482. Published online 2012 Jun 29. doi: 10.1118/1.4730287
.
The math and algorithms behind the calculations are presented in Information capacity measurements from Slanted edges: Equations and Algorithms.