Getting Started

System Requirements

Certifi Pedigree QP is compatible with Windows XP, Vista, and 7 platforms. For best performance, a 2 GHz or faster processor is recommended, and dual and quad-core CPUs can provide additional performance gains. Minimum recommended RAM is 2 GB for XP and 4 GB for Vista and 7, although Pedigree QP may work with less RAM depending on the size of the images being processed. As with all image processing applications, it is necessary to have sufficient disk capacity to store the input images and Certified output images.

Installing Certifi Pedigree QP

Pedigree QP requires Microsoft’s .NET Framework 3.5 to run. If it is not already installed on your computer, .NET 3.5 can be downloaded and installed from Microsoft’s website (http://www.microsoft.com/downloads).

After verifying that your computer has .NET 3.5 installed, Pedigree QP is installed by running (double-clicking) PedigreeQPInstaller.exe. It is highly recommended that you use the default installation directories to minimize any access permission issues and to simplify debugging in the event that a problem occurs. Installation can be done for a single user or for all user of the computer.

After installation is complete, there will be three Pedigree icons on the desktop:

· Pedigree QP Control Panel

· Monitor (Snapshot)

· View

Pedigree QP Control Panel is the main interface to Pedigree QP, and it is recommended that you start working with Pedigree QP by launching Control Panel.

Pedigree QP Control Panel Applications

The main interface to Certifi Pedigree QP is the Control Panel, which gives access to the commonly used applications in a quality-monitored image digitization workflow. The Control Panel applications include:

· Script Builder, for viewing images and analyzing image quality while interactively building and editing image processing scripts that can be saved for batch processing;

· Profile Builder, for building and editing image quality profiles that are used to determine if an image has acceptable quality;

· Engine, for performing batch processing and batch image quality analysis on a “hot” directory of images;

· Monitor (Real-Time), for monitoring the image quality results for image analyzed by Engine; and

· License, for checking how many credits remain on your account and to purchase more credits.

Launching Control Panel

Control Panel can be launched by one of these methods:

· Double-clicking on the PedigreeQPControlPanel icon on the desktop; or

· Choosing Start -> Programs -> Certifi ->Pedigree QP -> PedigreeQPControlPanel.

System Tray Icon

When Control Panel is launched, a Certifi Pedigree icon is also placed in the system tray.

System Tray

This icon serves as a way to access the Control Panel applications, and it also provides visual feedback on the image quality status by flashing green or red during batch processing with Engine (refer to Engine Status Indicators).

Control Panel Workflow: Single Image

After digitizing an image containing a Certifi Pedigree test target, Script Builder can be used to open the image file and analyze the quality of the image. Image processing can also be applied prior to the quality analysis using the integrated processing operations (refer to Using the Image Processing (IP) Editor), and the processing operations can be saved as a script for batch processing.

The quality measurements are compared against a specified quality profile that has been created by Profile Builder. It’s easy to change the quality profile in Script Builder so an image can be certified for different end uses such as high quality master files or lower quality derivative files.

Script Builder can also save the processed and analyzed image as a Certified file. One credit is deducted from the License account for each Certified file that is saved to disk. The License application in Control Panel can be accessed at any time to see the number of available credits. If there are no available credits, image processing and quality analysis can still be performed with Script Builder, but Certified files cannot be saved.

Control Panel Workflow: Batch Processing

In a batch workflow, the digitized images are placed in a “hot” directory that is continuously monitored by Engine. Images are processed by Engine in the order that they are loaded in the hot directory, and once an image is processed, it is moved to a processed image sub-directory under the hot directory.

IMPORTANT! The creation of the processed sub-directory and the transfer of images into it require the hot directory to have write permissions. This also means that images cannot be processed by Engine directly from read-only media such as a CD or DVD. In such cases, the images must first be transferred to a disk directory that has write permissions enabled.

Each image is automatically processed (if a processing script is supplied) and analyzed for quality, and the quality measurements are compared against a specified quality profile that has been created by Profile Builder.

Monitor (Real-Time) provides a continuously updated display of the quality results. The data displayed by Monitor is reset whenever Engine is stopped and restarted.

Certified image files can also be saved by Engine as it is processing the images in the hot directory. One credit is deducted from the License account for each Certified file that is saved to disk. The License application in Control Panel can be accessed at any time to see the number of available credits. If there are no available credits, Certified files cannot be saved when running Engine.

Launching Applications in Control Panel

The applications in Control Panel can be launched by one of these methods:

· Clicking one of the icons in Control Panel;

· Clicking Programs on the main menu in Control Panel and choosing an Open menu item; or

· Right-clicking on the Pedigree icon in the system tray and choosing an Open menu item.

Exiting Control Panel

Control Panel and all of its applications can be exited by one of these methods:

· Clicking Programs on the main menu in Control Panel and choosing Exit Control Panel; or

· Right-clicking on the Pedigree icon in the system tray and choosing Exit Control Panel.

IMPORTANT! Closing Control Panel by choosing the Close button in the upper right does not terminate Control Panel. This is because the Engine application in Control Panel can be used as a background process to continuously process and monitor a “hot” image directory. To prevent this background process from being turned off accidently, it is necessary to use one of the methods described above to explicitly terminate Control Panel and all of its applications.

Pedigree QP Stand-Alone Applications

Certifi Pedigree QP also includes two stand-alone applications to provide features that are useful outside of a production workflow:

· Monitor (Snapshot), for reviewing the image quality results for a directory of Certified images; and

· View, for viewing images and displaying the image quality measurements and the data integrity status of a Certified image.

Launching Monitor (Snapshot)

Monitor (Snapshot) can be launched by one of these methods:

· Double-clicking on the Monitor icon on the desktop; or

· Choosing Start -> Programs -> Certifi -> Pedigree QP -> Monitor.

Launching View

View can be launched by one of these methods:

· Double-clicking on the View icon on the desktop; or

· Choosing Start -> Programs -> Certifi -> Pedigree QP -> View.

Uninstalling Certifi Pedigree QP

Certifi Pedigree QP can be uninstalled by following these steps:

· Terminating Control Panel and any associated applications if the Pedigree icon is shown in the system tray;

· Choosing Start -> Settings -> Control Panel -> Add or Remove Programs; and

· Choosing Pedigree QP from the list of installed programs and clicking Remove.

Image Quality Metrics

Overall image quality is dependent upon a number of individual quality attributes:

· Is the sampling rate acceptable?

· Is the sharpness acceptable?

· Is the noise acceptable?

· Is the tonescale reproduction acceptable?

· Is the color reproduction acceptable?

· Is the exposure level acceptable?

· Is the exposure uniformity acceptable?

An image that meets all of these criteria is likely to be judged as having good overall quality.

To measure acceptability of each of these quality attributes, it is necessary to have one or more specific quality metrics associated with each attribute. For example, the attribute of noise can be measured using a standard deviation metric or a related metric such as peak signal-to-noise ratio (PSNR). The following sections describe the quality metrics that are used in Pedigree QP.

Sampling Rate

Sampling rate is the distance between adjacent pixels when mapped to the plane of the original content. Pedigree QP uses dots per inch (DPI) as the quality metric for the sampling rate. It’s measured by counting the number of pixels between the fiducial centers on the Certifi test targets, which are located a precise distance apart.

The sampling rate will vary with the camera-to-subject distance, so it’s important to have the test target in the same plane as the content being digitized for the most accurate DPI measurements.

The sampling rate will also vary over the image if the content is not co-planar with the camera sensor, which introduces a perspective distortion. The Certifi FC-1 target is useful in detecting this distortion as the maximum and minimum DPI can be reviewed as part of the quality analysis.

Sampling rate can be interpreted as a measure of the potential for sharpness. A higher DPI allows for smaller details to be seen, but only if the image is properly focused. A higher DPI also leads to larger file sizes, which can be a potential issue if large volumes of images are being processed or if storage space is limited.

Sharpness

Sharpness (often referred to as resolution) is a quality attribute that can be measured with a variety of different metrics. However, all of the sharpness metrics are derived from the spatial frequency response (SFR), which is also referred to as the modulation transfer function (MTF). Pedigree QP measures the SFR and also computes several quality metrics that are derived from the SFR.

In an image with high sharpness, edges will have a fairly abrupt transition from dark to light or vice versa. An image with low sharpness will have edges that transition over some distance (i.e., they’re blurred).

The SFR measures the degree of blurring. It is calculated by analyzing a particular type of edge known as a “slanted edge”. The FC-1 and FC-2 targets contain slanted edges of different orientations and contrast ratios.

Spatial Frequency Response (SFR)

The SFR is a curve that represents modulation (or amplitude) as a function of spatial frequency. Spatial frequency can be represented in normalized units of cycles/pixel, or in units of dots per inch if the sampling rate is known. While it’s beyond the scope of this User’s Guide to provide a detailed description of spatial frequency analysis, it is worthwhile to examine some idealized SFR plots to see what a desired response looks like.

If an edge has absolutely no blurring at all, the SFR would look like this:

However, this type of SFR is not really desirable in a digital system and it will never occur in practice because of the inherent blurring in optical systems and sensors.

Because of this inherent blurring, a typical SFR will look more like this:

The peak modulation of this SFR curve is 1.0 and it occurs at a spatial frequency of 0 cycles/pixel. The modulation drops as the frequency increases, reaching a value of around 0.10 at a spatial frequency of 0.5 cycles/pixel in this example.

The red line at 0.5 cycles/pixel represents the Nyquist frequency, which is an important value in a digital system. Ideally, the modulation at the Nyquist frequency would be in the range of 0.1 to 0.3. If the Nyquist frequency modulation is too low, the sharpness is less than it could be; if it’s too high, there’s the potential for aliasing, which is an undesirable image distortion.

Also shown in this plot are the spatial frequencies where the SFR modulation is 0.1 (10% SFR) and 0.5 (50% SFR). As already mentioned, the 10% SFR frequency should be close to the Nyquist frequency. The 10% SFR frequency is sometimes referred to as the “limiting resolution” because spatial frequencies with less than 10% modulation become imperceptible. The 50% SFR frequency is another way to characterize the sharpness of a system, and a higher spatial frequency for the 50% SFR point is generally desirable because it represents a sharper system.

The loss in sharpness that occurs with a blurred edge is clearly reflected in lower values for the 10% SFR and 50% SFR as shown in the following example. The 10% SFR point is now much less than the Nyquist frequency, and the spatial frequency for the 50% SFR point is also reduced significantly from the previous example.

It is common to apply some type of sharpening operation to an image to increase its sharpness. In fact, many cameras and scanner do this automatically. When sharpening is applied, the SFR will often look like this:

The peak modulation is now greater than 1.0 with this sharpened edge. While some sharpening can be desirable, it is possible to apply too much sharpening, which will further boost the peak modulation. Too much sharpening can result in a “halo” effect at edges as well as amplifying noise to unacceptable levels.

Finally, here’s an example of an SFR where the modulation at the Nyquist frequency is greater than the desired range of 0.1 to 0.3. In this case, there is the possibility for undesirable aliasing artifacts.

Effective DPI and Sampling Efficiency

The spatial frequency in an SFR plot can be converted easily to dots per inch if the sampling rate DPI is known. The relationship is that the Nyquist frequency at 0.5 cycles/pixel simply corresponds to the sampling rate DPI.

The 10% SFR frequency, represented as DPI, is referred to as the effective DPI. Equivalently, one can refer to the sampling efficiency, which is the ratio of the effective DPI to the actual sampling rate DPI:

In the following examples, we’ll use a sampling rate that corresponds to 400 DPI. For the typical edge example, the 10% SFR frequency is close to the Nyquist frequency, which means the 10% SFR frequency also corresponds to approximately 400 DPI. In this example, the effective DPI is also 400 DPI and the sampling efficiency is 100%.

In comparison, the effective DPI for the blurred edge example is 170 DPI, corresponding to a sampling efficiency of only 43%.

SFR and Orientation

The previous examples showed a single SFR curve. In imaging systems, it is important to measure the SFR in both the horizontal and vertical directions as they can be quite different.

In cameras, different horizontal and vertical SFR curves are typically the result of the interpolation that is used to create an RGB images from the subsampled color filter array (CFA) data. In scanners, there is also interpolation of the color channels, as well as mechanical differences in the fast and slow scan directions. External factors such as vibration or motion of the content during exposure can also affect the SFR in one or both directions.

SFR and Color Channels

There may also be significant differences in the SFR curves for the red, green, and blue channels. This is typically due to the color channel interpolation, but it may also result from a poor optical design that introduces chromatic aberrations.

As shown in the following example, it is common for the green channel to have the best SFR, but device manufacturers may sometimes include sharpening of the red and blur channels to compensate. In the example, a blacked dashed line that represents the luminance SFR has also been shown. Because the luminance channel is a combination of the RGB channels, its SFR curve will typically lie between the RGB curves. However, color misregistration also affects the luminance SFR and may further degrade it relative to the RGB curves.

Color Misregistration

Color misregistration refers to spatial offsets between the RGB color channels. It typically occurs because of improper interpolation of the subsampled RGB components in a camera or scanner sensor.

Color misregistration is reported as the pixel offsets for the red and blue channels relative to the green channel. Like the SFR, the amount of misregistration can be different in the horizontal and vertical directions. Misregistration is not derived directly from the SFR, but it’s computed at the same time using slanted edge data.

It is worthwhile to note that a system with matched SFR curves for the red, green, and blue channels may still exhibit significant misregistration. A well-designed imaging system should have minimal misregistration (< 0.5 pixel).

Slant edge misregistration3.tif

Noise

Noise is unwanted variations in the image data. The amount of noise is affected by numerous factors including the particular sensor that is used in a capture device, the amount of light falling on the sensor, how much gain is applied to the output of the sensor (e.g., the ISO speed), and the post-processing that is applied to the image (e.g., sharpening, noise reduction, contrast enhancement, etc.).

noise low density.jpg noise high density.jpg

Noise is typically measured using the standard deviation of the code values over uniform areas. Higher standard deviation values indicate more noise. Because the amount of noise is typically dependent on the signal level, it is useful to measure the noise as a function of the target density. An idealized plot of noise versus density for a camera with a CMOS sensor might look something like the following:

However, it’s important to note that the noise versus density curve can have almost any shape because of the widely different sensors and internal image processing operations that are used by equipment manufacturers.

Pedigree QP measures noise by calculating the standard deviation of the code values over uniform neutral patches. There are 15 neutral patches of different densities in the FC-2 and FC-2 test targets that are used for the noise calculations. Separate standard deviations are calculated for the red, green, and blue channels, as well as for the luminance.

Other noise metrics can be derived from the standard deviation, including peak signal-to-noise ratio (PSNR) and incremental signal-to-noise ratio (Incremental SNR).

Aim Noise Points

In addition to measuring the noise standard deviation values for all neutral patches, Pedigree QP calculates the noise standard deviations at three specific density levels, corresponding to the shadows, midtones, and highlights (density = 1.70, 0.80, and 0.07, respectively). These aim noise points are useful as summary quality metrics when determining if the noise meets a particular quality specification.

Color Reproduction

Color reproduction refers to the encoding of scene colors in a specified color space (e.g., sRGB, Adobe RGB, etc.). The goal is to encode colors so that they appear similar to the original scene colors when viewed on an appropriate display device, subject to the available colors of the display (i.e., the color gamut).

For proper color reproduction, it is essential to use a color-managed workflow. This means that the capture device should be characterized by its own ICC profile that has been obtained under the same lighting as will be used in production. This color profile can then be used to map the acquired image data into a well-defined color space such as sRGB.

All image processing should respect the color profile and propagate it appropriately throughout the workflow. If a display is used to evaluate subjectively the final color reproduction, it should also be characterized/calibrated with an ICC profile.

There are a number of color calibration packages (software and hardware) available in the marketplace to characterize the ICC profiles for capture and display devices.

Given a color-managed workflow with the proper ICC profiles, the issue of color reproduction in a digitization workflow then becomes a simpler job of tracking certain colors to make sure they don’t drift from the intended encoding. In Pedigree QP, this is done by measuring the neutrality deviations of the 15 neutral patches.

Neutrality deviations are quantified by converting the RGB code values of the neutral patches to an L*a*b* color space and calculating the delta E (ΔE) values for the chrominance components (ΔEab). The luminance component is ignored to prevent under or overexposures from biasing the results.

The purpose of delta E is to provide a metric that is perceptually uniform over the range of human vision. There are different ways to calculate delta E as the equations have been refined several times over the last 30 years, and Pedigree QP uses Delta E 2000 as it is the most recent metric.

IMPORTANT! If an image does not have an ICC profile associated with it, the ΔEab calculations are not performed. This is because it is impossible to convert the RGB values to L*a*b* if the meaning of the RGB values is unknown. Pedigree QP can add an ICC profile to an image or change an existing ICC profile by applying a color management operation in Script Builder (refer to Using the Image Processing (IP) Editor).

As a rule of thumb, two colors that differ by less than one ΔE are indistinguishable from one another. In practice, delta E values of 2 – 3 indicate good color reproduction. An example of ΔEab versus the target density is shown below. There is a slight increase in the ΔEab at the lowest densities, which would indicate the RGB tonescale curves are deviating from another in the highlight regions (refer to Tonescale Reproduction).

Aim Color Points

In addition to measuring the ΔEab values for all neutral patches, Pedigree QP calculates the ΔEab at three specific density levels, corresponding to the shadows, midtones, and highlights (density = 1.70, 0.80, and 0.07, respectively). These aim color points are useful as summary quality metrics when determining if the color reproduction meets a particular quality specification.

Tonescale Reproduction

Tonescale reproduction is the relationship between the light hitting a sensor and the resulting code values in the image. This relationship is fully described by the opto-electronic conversion function (OECF), which shows the code values versus the target reflectance, or equivalently, the code values versus the target densities. The following examples show what a typical OECF might look like when plotted against reflectance or density.

Pedigree QP measures the OECF using the 15 neutral patches in the FC-1 and FC-2 targets. OECF curves are calculated for each of the RGB color channels as well as for the luminance.

The measured OECF curves are dependent upon many factors, including the exposure level, the specific digital encoding used by the capture device, and any tonescale processing that is applied to the image (e.g., brightness/contrast and gamma adjustments).

For color images, it is necessary for the R,G,B OECF curves produced from neutral patches to be very similar to each other. This means that neutrals should have the same RGB code values at every density level, i.e., the neutrals should “track” over the tonescale range. If they don’t, the result is a color cast in the neutrals at one level or another.

One way of characterizing the OECF is to use an equation of the following form:

or equivalently,

The gamma parameter in these equations represents the gamma encoding that is used by a capture device. (It is also common to represent the gamma term as 1/gamma, which is actually the gamma of the intended display device.) With these types of equations, the OECF can be characterized by the three parameters: Gamma, Gain, and Offset. In addition to measuring the OECF, Pedigree QP also computes the best-fit gamma, gain, and offset terms.

An example of changing the gamma is shown below, where a gamma = 0.45 represents a typical OECF (corresponding to a display gamma of 2.2) and a gamma = 1.0 represents a linear OECF, which would be appropriate for a raw image file.

The gain term acts essentially like a contrast adjustment, while the offset term is similar to a brightness adjustment. Typical settings would be gain = 1.0 and offset = 0, but good tone reproduction can be obtained with other values if they are chosen carefully.

Exposure

Proper exposure means that the each reflectance (or density) value in a scene is mapped to the proper code value in the digital image. Determining if the proper exposure was achieved essentially amounts to defining an aim OECF curve and then comparing the measured OECF to the aim OECF. An example of the change in the OECF for overexposure and underexposure is shown below.

Aim Exposure Points

In addition to measuring the full OECF for the RGB and luminance channels, Pedigree QP calculates the exposure code values at three specific density levels, corresponding to the shadows, midtones, and highlights (density = 1.70, 0.80, and 0.07, respectively). These aim exposure points are useful as summary quality metrics when determining if the exposure meets a particular quality specification.

These aim exposure points can also verify the proper tonescale reproduction when used in conjunction with the measured gamma parameter for the OECF curve. A system that has the proper aim exposure points and also has the proper gamma parameter will produce an image that meets the desired aim OECF over the entire range of densities.

Clipping

Clipping can occur when an image is under or overexposed, or has been processed with too much brightness or contrast adjustment. Under these conditions, the shadow regions of the image may be set to the minimum code value in all color channels (0, 0, 0) or the highlight regions may be set to the maximum code value (255, 255, 255). This is sometimes termed as “crushing” the shadows or “blowing out” the highlights.

Pedigree QP measures clipping by computing the percentage of pixels that have a value of (0, 0, 0) and the percentage of pixels that have a value of (255, 255, 255). This quality metric is computed from the entire image, not just from the target data, because the image content may include a larger range of densities than are contained in the target.

Exposure Uniformity

Exposure variations across the image plane can occur because of illumination nonuniformity or light falloff in lens. The exposure uniformity can be measured by digitizing a uniform target and measuring the variations in code values across the image.

Pedigree QP uses a finely spaced sampling grid across the FC-1 and FC-2 targets to measure the code value variations in the uniform background regions.

The code value variations are converted into exposure variations by using the OECF. The uniformity deviation is reported as the max-to-min percent deviation in the exposure level. Pedigree QP measures the exposure uniformity for each of the RGB color channels, as well as for luminance.

The uniformity measurements are mostly of benefit when used with the FC-1 target because of its larger area coverage. Also, care must be taken with the smaller FC-2 target as adjacent content may cast shadows on the target, resulting in erroneous uniformity measurements.

The finely spaced sampling that is used in the Pedigree QP uniformity analysis allows for the creation of a uniformity map that represents the exposure variations at the image sensor plane. In constructing this uniformity map, if any sample points fall on an area other than the uniform background, the samples are ignored and instead are interpolated from surrounding values.

The uniformity map can be a useful aid in adjusting the lighting for a camera-capture setup.

Target Interval

In some applications, a test target cannot be included in every image. In such cases, the quality metrics from the last test target are propagated into subsequently acquired images until another test target is encountered.

Pedigree QP keeps track of the time interval and number of images that have been analyzed since the last target was encountered. While these two metrics are not image quality metrics in the traditional sense, they are possible indicators of quality issues because the performance of digitization equipment can drift over time and as more image are processed.

Data Representations for Image Quality Metrics

XMP Quality Metadata

The primary data representation for the image quality metrics in a Certified file is Adobe’s Extensible Metadata Platform (XMP). Serialized XMP data is represented using a subset of the W3C Resource Description Framework (RDF) data model, which uses an XML-based format.

The benefit of an XMP representation is that Adobe has already defined the protocols for embedding XMP metadata into the most common image formats, including TIFF, JPEG, JPEG 2000, and PDF. Most modern imaging applications will automatically propagate XMP metadata, so the Certifi image quality metrics will be carried along wherever an image might be sent. In addition, the XML formatting of XMP metadata makes it human-readable without any special processing.

However, the downside of XMP is that the RDF data model is somewhat verbose, leading to overhead in the Certified image file. To minimize this overhead, a subset of the full quality metrics that are measured by Pedigree QP is stored in the XMP metadata (refer to XMP Quality Metadata Summary).

Deep-Dive Quality Metadata

In addition to the XMP metadata that is embedded in the Certified image file, Pedigree QP also creates a more detailed quality metadata file (the “deep-dive” metadata) that is associated with the Certified file. The deep-dive metadata file includes all of the quality metrics that are measured by Pedigree QP and is represented using XML. The deep-dive file is stored in a subdirectory (“..\deepIq”) of the directory that contains the Certified file and is linked by an element in the embedded XMP metadata.

The deep-dive XML data will be available as long as the deepIq subdirectory is included when a directory of Certified images is moved or renamed.

Summary of Certifi Image Quality Metrics

XMP Quality Metadata Summary

The following table lists the specific quality metrics that are stored in the XMP metadata of a Certified file.

Quality Attribute	Certifi XMP Quality Metrics	Additional Details
Sampling rate	Min/Max DPI	Horizontal and vertical
Sharpness	Effective DPI SFR 10% SFR 50% SFR Peak Modulation SFR Nyquist Modulation Misregistration	Horizontal and vertical for RGB and luminance
Noise	Standard deviation	Three density aim points for RGB and luminance
Color reproduction	ΔEab	Three density aim points
Tonescale reproduction	Gamma Gain Offset	RGB and luminance
Exposure level	Output CV % Clipping	Three density aim points for RGB and luminance Shadows/highlights
Exposure nonuniformity	% Deviation (min to max)	RGB and luminance
Target Interval	Number of images since last target Time since last target

Table 1. Quality metrics in the embedded XMP metadata.

Deep-Dive Quality Metadata Summary

The following table lists the specific quality metrics that are stored in the separate deep-dive metadata file.

Quality Attribute	Certifi Deep-Dive Metrics	Additional Details
Sampling rate	9 DPI measurements across field of target (FC-1 target only)	Horizontal and vertical
Sharpness	SFR (complete data) Effective DPI SFR 10% SFR 50% SFR Peak Modulation SFR Nyquist Modulation Misregistration	Horizontal and vertical for RGB and luminance
Noise	Standard deviation Peak SNR Incremental SNR	15 density levels for RGB and luminance
Color reproduction	ΔEab	15 density levels
Tonescale reproduction	OECF (complete data) Gamma Gain Offset	15 density levels RGB and luminance RGB and luminance
Exposure level	OECF (complete data) % Clipping	15 density levels RGB and luminance Shadows/highlights
Exposure nonuniformity	Nonuniformity map % Deviation (min to max)	RGB and luminance
Target Interval	Number of images since last target Time since last target
Geometry	Perspective transformation parameters (FC-1 target only)

Table 2. Quality metrics in the deep-dive XML file.

Using Script Builder

Certifi Pedigree^® ScriptBuilder is used to view images and analyze their quality in real-time. The quality results are compared against a specified quality profile to determine if the quality is acceptable for a given application.

Viewing Images

Supported file formats

Pedigree QP supports TIFF and JPEG (baseline) formats for input images. TIFF images can be 8 or 16-bits/color channel, while baseline JPEG is always 8 bits/color channel. Binary images (1 bit/pixel) are not supported because the limited bit depth renders most of the quality metrics as useless.

There is also limited support for PDF/A-1b input images in that some, but not all, PDF/A-1b images can be read.

Loading an image from a directory

An image is loaded for viewing by:

· Choosing File -> Open Image from the main menu, or

· Clicking on the Open Image File button on the main toolbar.

A file browser dialog will appear, allowing you to navigate to a specific directory. Once in the desired directory, an image file can be selected by one of these methods:

· Double-clicking on the image filename, or

· Clicking on the image filename and then clicking Open.

The filename of the currently displayed image is always shown in the title bar.

The image is always displayed at a magnification that allows it to fit within the image window unless this option is turned off (refer to Setting Script Builder Options).

Navigating between images within a directory

Once an image has been loaded, you can navigate among the other images in the same directory by using the Next Image and Previous Image buttons on the main toolbar.

If the current image is the first image in a directory, the Previous Image button will be disabled. Likewise, if the current image is the last image in a directory, the Next Image button will be disabled.

Navigating between pages in a multipage image

If the current image contains multiple pages, you can navigate among the pages by using the Next Page and Previous Page buttons on the main toolbar.

If the current page is the first page in a multipage image, the Previous Page button will be disabled. Likewise, if the current page is the last page, the Next Page button will be disabled.

The total number of pages and the currently displayed page are shown in title bar next to the image filename.

Using the zoom controls

The Increase Magnification button on the main tool bar allows you to zoom in quickly to examine fine details.

The Decrease Magnification button allows you to zoom out quickly to see a larger area of the image.

The Specify Magnification button allows you to choose a specific magnification percentage or fit the image to the display window from its drop-down menu.

When the cursor is inside the image area, the mouse wheel can also be used to quickly zoom in and out.

When used in conjunction with the selection box (refer to Using the selection box), these zoom controls allow you to zoom in and out on a selected region. A selection box must first be drawn around the desired region and the zoom controls will then use the center of the selection region as the focus point.

Displaying an alignment grid overlay

The Display Alignment Grid button allows you to display a grid of horizontal and vertical lines on the image to check for alignment of image features. The grid can be turned on/off at any time by clicking on the button.

Using the selection box

The selection box allows you to select a desired region in the image. If no selection is made, the entire image is selected by default.

A region is selected by:

· Holding the left mouse button down anywhere in the displayed image, and

· Dragging the mouse to another point in image.

These steps will draw a rectangular selection box in dashed lines on the image. The selection box is removed by double-clicking anywhere in the displayed image. This operation will select the entire image.

The selection box size and coordinates (upper-left and lower right x,y values) are displayed in the selection box panel. The units for the size and coordinates can be changed by using the drop-down menu in the selection box panel.

When the units are changed to physical distances (inches, centimeters), it is necessary to know the DPI. If a Certifi target is present in the image, the DPI is automatically calculated and used to convert pixel distances into inches or centimeters. If there is no target, the DPI from the input image metadata is used if available. Otherwise, a default DPI is used (refer to Image metadata options). The DPI source is indicated in the selection box panel.

The selection box can be changed by one of these methods:

· Drawing a new selection box;

· Moving the cursor to one of the edges or corners of an existing selection box until a edge or corner cursor appears, holding the left mouse button down, and dragging the mouse to a new location;

· Moving the cursor to the center of the selection box until a crosshairs cursor appears, holding the left mouse button down, and dragging the mouse to a new location; or

· Double-clicking anywhere in the displayed image to select the entire image.

Reading individual pixel values

Individual RGB pixel values are displayed real-time in the pixel value panel as the mouse is moved over the image. The x,y pixel coordinates can be in pixels, inches, or centimeters (refer to Using the selection box for changing units).

It may be necessary to zoom in to accurately capture the pixel value at a desired location in the image.

Displaying image statistics (histogram)

The Display Image Statistics button allows you to view the histogram and statistics for the current image or for a selected region using the selection box. The Image Statistics dialog can be turned on/off at any time by clicking on the button. The histogram and statistics are updated automatically when a new image is displayed or when the selection box is changed.

Displaying image metadata

The metadata for the current image is displayed by:

· Choosing Image -> Image Information from the main menu.

Using the Image Processing (IP) Editor

IP editor panel

The IP editor panel provides access to image processing operations that can be applied to an image prior to the image quality analysis and the saving of a Certified file. Pedigree QP includes a suite of basic and advanced image processing operations

The IP editor toolbar allows you to add/delete image processing operations to create an image processing (IP) script, which can then be saved to a file. Existing IP scripts can also be opened from the IP editor toolbar.

The list of processing operations in the IP editor always starts with two operations: Original and Image Quality Analysis. Additional processing operations are added between Original and Image Quality Analysis. The image that is displayed in the image window always corresponds to the operation that is highlighted in the processing operations list.

Refer to Appendix A: Image Processing Operations for descriptions of the available image processing operations and how to use them.

Adding an image processing operation

An image processing operation is added by clicking on the Add Step button on the IP editor toolbar and selecting one of the available operations from the drop-down menu. Once an operation is selected, the interactive dialog for the selected operation is opened. Clicking “OK” on the dialog will add the operation to the list of processing operations in the IP editor.

It is important to note that a new image processing operation can only be added after the Original step and before the Image Quality Analysis step. The operations can be added to image processing script in any order, except for Color Management, which must be added as the first operation because it potentially affects all of the other processing.

Importantly, the software is designed so it is possible to remove the test target with one of the crop operations and still perform the Image Quality Analysis. This is accomplished by maintaining a copy of the test target that is processed through the same operations as the main image.

Selecting an image processing operation

An image processing operation is selected by left-clicking on the operation name in the operations list. Selecting an operation will cause the IP editor to process an image by applying every processing operation up to and including the selected step. The resulting processed image is displayed in the image window.

Deleting an image processing operation

An image processing operation is deleted by:

· Right-clicking on the operation to open the IP operations menu; and

· Selecting Delete Step from the menu.

The selected operation will be removed from the operations list. It is not possible to delete the Original step or the Image Quality Analysis step.

You can also delete all image processing operations (with the exception of Original or Image Quality Analysis) by clicking on the Delete Current Script button on the IP editor toolbar.

Editing an existing image processing operation

An existing image processing operation is edited by:

· Right-clicking on the operation to open the IP operations menu; and

· Selecting Edit Step from the menu.

The interactive dialog for the selected operation will then be opened. Clicking “OK” on the dialog will apply any edited changes, while clicking “Cancel” will discard any changes. It is not possible to edit the Original step or the Image Quality Analysis step.

Saving an image processing script

The Save Current Script button on the IP editor toolbar allows you to save the current IP script to a file.

Opening an existing image processing script

The Open Existing Script button on the IP editor toolbar allows you to import an existing IP script into the IP editor. The imported IP script can then be edited like any other script.

Analyzing Image Quality

Target identification

The process of analyzing an image for quality begins as soon as the image is loaded. The test target (if present) is identified automatically and displayed in the status bar at the bottom of Pedigree QP.

Quality analysis

The IP editor panel highlights the Original step as an image is loaded and analyzed for the presence of a test target. The Image Quality Analysis step is then highlighted as the automated quality analysis begins.

If additional IP operations have been added (refer to Adding an image processing operation), the operations will be applied to the original image prior to the quality analysis. Each operation name is highlighted as the operation is applied.

IQ panel

The Image Quality (IQ) panel displays the quality metrics that are contained in Certifi XMP metadata (refer to XMP Quality Metadata Summary).

IQ Panel.jpg

The image quality metrics are compared against the quality profile that is specified in the IQ panel. The profile is changed by using the drop-down menu for the IQ profile. Pedigree QP includes four pre-defined quality profiles (refer to Using Profile Builder for a description of these profiles and the details of each displayed quality metric).

IQ panel profile selection.jpg

The quality metrics in the IQ panel are color-coded to indicate if they meet the specifications in the selected quality profile (green for within spec; red for outside of spec). The profile “IQ_noLimits” has very broad limits for all quality metrics so that all metrics will always be within spec.

The status bar includes an overall quality indicator. If all quality metrics are within spec, the quality indicator will be “OK”. If any single quality metric is outside of its spec, the quality indicator will be “Not OK”.

Displaying deep-dive quality metrics from IQ panel

Each quality metric in the IQ panel is linked to its corresponding deep-dive information. The deep-dive quality metrics are displayed by double-clicking on the individual quality metrics.

In this deep-dive example, the full SFR information for the vertical direction is shown.

Setting Script Builder Options

Image processing options

Image processing options are changed by:

· Choosing Tools -> Options from the main menu, and

· Clicking on the Image Processing tab.

The “Always run processing script to end” option will automatically execute the entire IP script when an IP operation is added, edited, or deleted. The “Always fit image to window” option can be deselected if you want to maintain the same magnification when loading different images.

The Reduced Memory Options are useful when working with large images in the IP editor. The “Don’t store intermediate images in memory” option doesn’t save any images besides the original image and the current displayed image, which may result in additional processing time as you move between operations in the IP editor or when you add or delete an operation. The “Don’t store original image in memory” option should be used as a last resort if insufficient memory is a problem as it forces a complete image reload from disk, which can be very time-consuming when dealing with large images.

Image metadata options

Image metadata options are changed by:

· Choosing Tools -> Options from the main menu, and

· Clicking on the Image Metadata tab.

The “Include input image metadata” option allows you to specify the types of input metadata (if available in the input image) that are propagated to the Certified image file.

The “DPI” option allows you to specify the source of DPI information. It is recommended that “Use DPI measured from target” be selected to take advantage of the precise DPI measurements available if a Certifi test target is used. If this option is selected and no test target is present, the DPI from the input image metadata is used. If there is no DPI metadata in the input image, the default DPI is used.

Similarly, if the “Use DPI from input image metadata” option is selected and there is no DPI metadata in the input image, the default DPI is used.

If the “Always use default DPI” is selected, the specified DPI value is always used regardless of whether a Certifi test target is present or the input image contains DPI metadata.

For convenience, the “Set same options in Engine” can be checked to propagate these Script Builder option settings to the options available under Pedigree Engine.

Document metadata options

Document metadata options for PDF output files are changed by:

· Choosing Tools -> Options from the main menu, and

· Clicking on the Document Metadata tab.

The Title, Author, Subject, Keywords, and Creator are entered by typing in the text boxes.

For convenience, the “Set same options in Engine” can be checked to propagate these Script Builder option settings to the options available under Pedigree Engine.

Compression options

Compression options for TIFF and PDF output files are changed by:

· Choosing Tools -> Options from the main menu, and

· Clicking on the Compression tab.

For TIFF output images, the compression may be set to None, LZW, or Zip/Deflate.

For PDF output images, the compression may be set to None or Zip/Deflate.

For convenience, the “Set same options in Engine” can be checked to propagate these Script Builder option settings to the options available under Pedigree Engine.

Saving Certified Files

A Certified image is saved to disk by:

· Choosing File -> Save Certified Image from the main menu, or

· Clicking on the Save Image File button on the main toolbar.

A file browser dialog will appear, allowing you to navigate to a specific directory and save the file. The output image format is specified in the file browser.

One credit is deducted from your License account for each Certified file that is saved to disk (refer to Using License).

Supported formats

Pedigree QP supports TIFF and PDF/A-1b formats for output images. TIFF images can be 8 or 16-bits/color channel, while PDF/A-1b always 8 bits/color channel.

Using Engine

Certifi Pedigree^® Engine allows you to perform batch quality analysis on an image directory. The directory is “hot”, so images placed in the directory are automatically added for analysis. Every image is automatically analyzed for image quality against a selected quality profile, and the quality results can be viewed using Certifi Pedigree^® Monitor (Real-Time) (refer to Using Monitor (Real-Time)).

Setting the Input Image Directory

The input directory is set by clicking the directory browser button next to Input Image Directory. A folder browser dialog will appear, allowing you to select a specific directory.

Setting the Certified Image Directory

The Certified output directory is set by clicking the directory browser button next to Certified Image Directory. A folder browser dialog will appear, allowing you to select a specific directory.

Unchecking the “Save Certified image files” option allows Engine to analyze images for quality without actually creating Certified files from them. The quality results can still be viewed using Monitor (Real-Time). When Engine is stopped and restarted, any previous quality results are discarded.

One credit is deducted from your License account for each Certified file that is saved to disk (refer to Using License).

Setting the Image Quality Profile

The image quality profile is set using the drop-down menu next to Image Quality Profile. All available profiles are displayed. Additional quality profiles can be built using Profile Builder (refer to Using Profile Builder).

Setting Engine Options

Image format options

Image format options are changed by:

· Choosing Tools -> Options from the main menu, and

· Clicking on the Image Format tab.

For TIFF output images, the compression may be set to None, LZW, or Zip/Deflate.

For PDF output images, the compression may be set to None or Zip/Deflate.

For convenience, the “Set same options in Script Builder” can be checked to propagate these Engine option settings to the options available under Pedigree Script Builder.

Image filename options

Image filename options are changed by:

· Choosing Tools -> Options from the main menu, and

· Clicking on the Image Filename tab.

If “Use input filename as output filename” is selected, the input image filename is used for the Certified image filename.

If “Use auto indexing for output filename” is selected, the Certified image filename is generated from a prefix filename and an index that is automatically incremented for each image that is saved. You can customize the auto indexing by specifying a starting index and whether odd-only or even-only indexing should be used.

The indexing can be reset to the starting index when Engine is started/stopped, or the indexing can be maintained across different runs of Engine.

A time stamp can also be added to the filename.

Image metadata options

Image metadata options are changed by:

· Choosing Tools -> Options from the main menu, and

· Clicking on the Image Metadata tab.

The “Include input image metadata” option allows you to specify the types of input metadata (if available in the input image) that will be propagated to the Certified image file.

Similarly, if the “Use DPI from input image metadata” option is selected, and no DPI information in the input image, the default DPI is used.

If the “Always use default DPI” is selected, the specified DPI value is always used regardless of whether a Certifi test target is present or the input image contains DPI metadata.

For convenience, the “Set same options in Script Builder” can be checked to propagate these Engine option settings to the options available under Pedigree Script Builder.

Document metadata options

Document metadata options for PDF output files are changed by:

· Choosing Tools -> Options from the main menu, and

· Clicking on the Document Metadata tab.

The Title, Author, Subject, Keywords, and Creator are entered by typing in the text boxes.

For convenience, the “Set same options in Script Builder” can be checked to propagate these Engine option settings to the options available under Pedigree Script Builder.

Starting/Stopping Engine

Engine is started by clicking the Start Engine button. When Engine is started, existing images in the input image directory are processed, starting with the earliest timestamp. All existing images are processed before any images that were added after Engine was started.

Engine is stopped clicking the Stop Engine button.

The results of the batch quality analysis are viewed using Monitor (Real-Time). When Engine is stopped, the quality results are maintained in Monitor until Engine is restarted.

Engine Status Indicators

When Engine is running, the Engine icon in Control Panel will display a moving yellow segment.

When Engine is running, the Monitor icon in Control Panel will alternate flash green or red to indicate the current quality status. A green indicator means that all images have met the quality specifications in the specified quality profile, while a red indicator means that at least one image has failed to meet the quality specifications.

The Certifi system tray icon also flashes red or green while Engine is running to indicate the quality status.

When Engine is stopped, the Certifi system tray icon will stop flashing and will remain red or green to indicate the quality status.

Appendix A: Image Processing Operations

Pedigree QP includes the following image processing operations:

· Rotate

· Auto Deskew

· Crop

· Auto Page Crop

· Auto Target Crop

· Add Border

· Resize

· Sharpen

· Unsharp Mask

· Despeckle

· Blur

· Brightness/Contrast

· Gamma

· Grayscale Conversion

· White Balance

· Color Management

The following sections provide an overview of each of these operations.

Rotate

The Rotate operation allows you to rotate an image by an arbitrary angle or by multiples of 90°. A positive rotation angle rotates an image clockwise, and a negative rotation angle rotates an image counter-clockwise.

The Rotate dialog includes options for the interpolation method (nearest-neighbor, bilinear, or cubic), with bilinear or cubic interpolation recommended for most applications.

Because a rotate operation introduces new regions around the image, it is necessary to fill those regions. The default fill color is black, but the Rotate dialog allows you to specify a different fill color or the fill color can be automatically determined using the average color at the edges of the original image. The automatic fill color option is particularly useful when the document or page has been placed on a background with a relatively uniform color, where the fill color can provide a fairly seamless match to the background.

The Display Alignment Grid option on Script Builder’s main tool bar (refer to Displaying an alignment grid overlay) provides a convenient way to check if various structures are aligned with the horizontal and/or vertical axes during a Rotate operation.

Auto Deskew

The Auto Deskew operation allows you to automatically correct unwanted rotation in a document or book page. The Auto Deskew operation works by analyzing all of the image contents, not just the page edges, so it can correct for text that is not aligned with the page edges. With hand-set type, there may be blocks of text that are all oriented slightly differently, and the Auto Deskew operation optimizes the deskew correction over all of the different text regions. Certifi test targets are automatically ignored during the Auto Deskew operation and hence their presence doesn’t affect the result.

The Auto Deskew dialog includes a slider control for the deskew angle search range, up to 45°, but typical skew amounts are usually less than 5-10°. There is also a slider control for the “Gutter”, which is the outside region of the image to be ignored while computing the optimal deskew angle. Larger gutter values focus more on the center of the image, while smaller values include more of the image contents, potentially including the page edges.

A deskew operation introduces new regions around the image because of the rotation, and it is necessary to fill those regions with a color value. The default fill color is black, but the Auto Deskew dialog allows you to specify a different fill color or the fill color can be automatically determined using the average color at the edges of the original image. The automatic fill color is particularly useful when the document or page has been placed on a background with a relatively uniform color where the fill color can provide a fairly seamless match to the background.

The Auto Deskew dialog also includes an option to crop the deskewed image to the original image size. This is useful when the amount of skew is small and it is desired that all images maintain a constant size.

Crop

The Crop operation allows you to crop a rectangular region of an image. The crop region can be specified by drawing a selection box on the image (refer to Using the selection box), and the selection box can be modified at any time while using the Crop dialog. Alternatively, the crop region can be specified using the crop window location (Start X, Start Y) and window size (Width, Height) on the Crop dialog.

The Crop dialog also includes an option to “Optimize crop window position automatically”. This option adjusts the crop window position (not its size) to center the page text in the crop window. This option should only be used when the Crop operation is preceded by Auto Deskew and Auto Page Crop operations, which will align the text and isolate it for the final cropping.

Auto Page Crop

The Auto Page Crop operation allows you to automatically crop a cut sheet or bound book page to the page boundaries, with optional margins allowing you to add or remove extra distance from the detected page boundaries. If a target is present in the image, it is ignored by the Auto Page Crop operation. The Auto Page Crop algorithm is quite advanced, and it doesn’t require the content to be against a high contrast background to detect the page boundaries.

The Auto Page Crop dialog is more complex than the other image processing dialogs in Pedigree QP, and there is a learning curve involved to achieve the desired results. This is because there are almost infinite combinations of page edge structure and squareness, page-to-background contrast, proximity of text and other content to the page edges, crease structures in bound books, capture equipment and associated lighting, etc., and all of which can affect the final result. However, once the settings are determined for a particular type of content, the Auto Page Crop operation can reliably extract the pages for that type of content. The following is a brief description of how to use the Auto Page Crop dialog, and it is recommended that you experiment with different types of content to get a better understanding of how to use this operation.

We’ll start first with an example of a cut sheet. In this case, using the Auto Page Crop dialog is relatively straightforward. When the dialog is shown, the image will be displayed with a selection box that highlights the crop region that was automatically detected. The dialog tab labeled “Margins” includes settings for the top, bottom, left, and right margins that are added/subtracted to the initial selection box. Positive margin values add additional space around the selection box, while negative margin values bring the selection in towards the center of the page. Clicking the “Reset” button will reset all of the margins to zero.

The slider controls for “Page-to-surround contrast” and “Page edge structure” allow for fine-tuning of the detected page edges depending upon the nature of the content and the background. The “Target shadow size” control can be used for cases where a test target is very close to the edge of the page, and the target has introduced a shadow on the page edge because of the lighting that was used.

Now let’s look at an example of a bound book page. In this example, there is no left surround, so the option labeled “No left surround” should be selected. This extends the selection box over to the extreme left side of the image. When any of the “No surround” options are selected, a new tab labeled “Crease” appears on the Auto Page Crop dialog. The Crease tab provides the tools to automatically detect the crease region and crop the page to the crease location as described on the next page.

On the Crease tab, the option labeled “Crop to crease” option must be selected. This option will highlight a crease search region on the displayed image. The crease search region is adjusted using the “Crease search range” sliders. It is not necessary to tightly restrict the search range, and in fact, you should leave some margin in the search range so positioning differences from page to page won’t cause the actual crease to go outside the search range.

At this point, the detected crease should be highlighted as part of the page selection box. You can choose additional options to refine the crease location (“Full crease”, “Split crease”, or “No crease”), along with a crease margin if desired. There is also an additional option labeled “Saturated crease”, which can be beneficial when the text or other content bleeds up to the crease.

Auto Target Crop

The Auto Target Crop operation allows you to automatically crop out an FC-2 target from an image. When the Auto Target Crop operation is selected, a selection box is shown on the displayed image to indicate the region that will be retained after cropping.

Pedigree QP software is designed to ensure that a full quality analysis is performed even when the target is removed from the image. The only assumption when using automatic target cropping is that the FC-2 target is located to one side (left, right, top, or bottom) of the image. If an image contains an FC-1 target, no cropping is performed because the FC-1 target usually fills the image frame.

The Auto Target Crop dialog is simple to use. The only option is to specify an extra margin around the target, but this is usually not needed.

Add Border

The Add Border operation allows you to add a border around an image, with settings for the border dimensions on the top/bottom and left/right sides and the border color. Options allow the bottom border dimensions to be locked to same size as the top border dimensions and likewise for the right and left border dimensions. Clicking the “Reset” button will reset all of the borders to zero.

There is also an option to automatically determine the border color using the average color at the edges of the image. The automatic border color is useful when the document or page has been placed on a background with relatively uniform color or when the document or page has a consistent color around its edges.

Resize

The Resize operation allows you to reduce or enlarge an image, either by specifying the width and height in pixels or a magnification percentage. The aspect ratio is always maintained at 1:1, so if you specify the width, the height is automatically computed and vice versa. Any magnification less than 100% (i.e., a reduction) automatically includes aggregation to minimize aliasing artifacts.

The Resize dialog includes options for the interpolation method (nearest-neighbor, bilinear, or cubic), with bilinear or cubic interpolation recommended for most applications.

Sharpen

The Sharpen operation allows you to enhance the fine detail in an image. The Sharpen dialog is very simple, with only a control for the sharpening percentage (from 0 to 500%). A higher value for the sharpening percentage produces sharper images, but it also increases any noise that is present in the image.

Unsharp Mask

The Unsharp Mask operation is a type of sharpening operation that gives more control over the sharpening process than does the Sharpen operation. The Unsharp Mask dialog has two slider controls, one for the radius of the mask (in pixels) and the other for the sharpening percentage (from 0 to 500%). A higher value for the mask radius or sharpening percentage increases the amount of sharpening. However, increasing the amount of sharpening will also increase any noise that is present in the image.

Despeckle

The Despeckle operation allows you to perform noise cleaning while minimizing degradations of desired content. The Threshold control in the Despeckle dialog determines how much original content is retained (from 0 to 100%) and the Iterations control determines how much smoothing is performed (from 1 to 5 iterations). In practice, there is very little additional benefit beyond 2 iterations.

Blur

The Blur operation is a simple smoothing process that can be used to reduce the noise in an image and/or soften edges that are oversharpened.

Brightness/Contrast

The Brightness/Contrast operation allows you adjust the overall brightness and contrast of an image. The Brightness slider control adds or subtracts a specified number of code values from every pixel, while the Contrast slider control scales the code values by a multiplier, centered around the midpoint code value (e.g., 128 for an 8-bit image).

The Brightness/Contrast dialog includes a graphical display of the relationship between the original code values and the output code values for the Brightness/Contrast operation. The original code values are represented by the horizontal axis and the output code values are represented by the vertical axis.

Gamma

The Gamma operation allows you adjust the gamma of an image. The Gamma slider control essentially applies an exponential function to the image code values, which causes the tonescale to bend so that code values become darker or lighter in a nonlinear way.

The Gamma dialog includes a graphical display of the relationship between the original code values and the output code values for the Gamma operation. The original code values are represented by the horizontal axis and the output code values are represented by the vertical axis.

Grayscale Conversion

The Grayscale Conversion operation allows you to convert an RGB image to grayscale (e.g., 24-bit RGB to 8-bit grayscale).

The Grayscale Conversion dialog includes an option to “Only convert to grayscale when low chrominance”, which allows you to set a chrominance threshold. With this option, an image with a chrominance (“colorfulness”) value greater than the threshold will not be converted to grayscale. This is useful when digitizing pages or books that are primarily black-and-white but some pages occasionally contain localized color areas such as stamps or seals. Applying this option will keep the color information for those pages that exceed the chrominance threshold, but convert the other pages to grayscale to save storage space.

White Balance

The White Balance operation allows you to balance the RGB channels at a single point on the tonescale curve so that white neutral colors will be reproduced properly.

If a Certifi test target is included in the image, the White Balance operation can automatically detect the target and set the white balance using the brightest patch that is not clipped (i.e., overexposed). This white balance is maintained until a new target is detected.

The White Balance dialog also allows you to select a manual option, where the white balance is set by drawing a selection box in the desired area to be used as the balance point. This manual white balance setting is then used to balance all other images.

Color Management

The Color Management operation allows you to specify the ICC profile associated with an image at various points in the processing chain. When the ICC profile for an image is changed within Pedigree QP, the code values of the image are modified to reflect the new ICC profile. An ICC profile is commonly referred to in terms of its color space, such as sRGB or Adobe RGB, but an ICC profile also contains other information including the tonescale, usually specified by the display gamma (e.g., 1.8 or 2.2).

There are four ICC profiles that can be set with the Color Management operation:

· Input profile: An input image may already have an embedded ICC profile, but the Color Management operation can also be used to attach an ICC profile to an image if there isn’t an embedded profile.

· Working profile: The working profile represents the color space in which the image processing operations will take place. The working profile may be the same as the input profile, but it is also common to use a wider gamut color space (e.g., Adobe RGB) for the working image. It is also possible to increase the bit depth of the working image (e.g., 8 bits -> 16 bits) to minimize the potential for quantization artifacts when using a wider gamut color space.

· Output profile: The output profile represents the ICC profile embedded in the processed output image. The output profile may be the same as the input profile and/or the working profile, and the bit depth can also be changed for the output image.

· Display profile: The display profile represents the expected ICC profile for the image data that is being displayed with Pedigree QP. Typically, this is sRGB. The display profile is never actually embedded in the image and only serves the role of a virtual profile during display. Failure to set the display profile may result in inaccurate colors in the displayed image, for example, displaying an image with an embedded Adobe RGB profile on a monitor that is expecting sRGB data.

The Color Management dialog is set up to allow each of these profiles to be easily modified. ICC profiles are set separately for color images and monochrome images by selecting the “Color” or “Grayscale” tab on the dialog. Pedigree QP includes common ICC profiles, such as sRGB and Adobe RGB for color images and Gray Gamma 1.8 and Gray Gamma 2.2 for monochrome images, but you can use any other valid ICC profile.

The default settings for the Color Management dialog are:

· Input profile: Use the embedded profile if one exists. If there isn’t an embedded profile, use sRGB for color images and Gray Gamma 2.2 for monochrome images.

· Working profile: Use the same profile as the input profile.

· Output profile: Use the same profile as the working (input) profile.

· Display profile: Use sRGB for color images and Gray Gamma 2.2 for monochrome images.

Launching Control Panel

System Tray Icon

Control Panel Workflow: Single Image

Control Panel Workflow: Batch Processing

Launching Applications in Control Panel

Exiting Control Panel

Launching Monitor (Snapshot)

Launching View

Spatial Frequency Response (SFR)

Effective DPI and Sampling Efficiency

SFR and Orientation

SFR and Color Channels

Color Misregistration

Aim Noise Points

Aim Color Points

Tonescale Reproduction

Aim Exposure Points

Clipping

XMP Quality Metadata

Deep-Dive Quality Metadata

Summary of Certifi Image Quality Metrics

XMP Quality Metadata Summary

Deep-Dive Quality Metadata Summary

Using Script Builder

Supported file formats

Loading an image from a directory

Navigating between images within a directory

Navigating between pages in a multipage image

Using the zoom controls

Displaying an alignment grid overlay

Using the selection box

Reading individual pixel values

Displaying image statistics (histogram)

Displaying image metadata

Using the Image Processing (IP) Editor

IP editor panel

Adding an image processing operation

Selecting an image processing operation

Deleting an image processing operation

Editing an existing image processing operation

Saving an image processing script

Opening an existing image processing script

Target identification

Quality analysis

IQ panel

Displaying deep-dive quality metrics from IQ panel

Setting Script Builder Options

Image processing options

Image metadata options

Document metadata options

Compression options

Supported formats

Using Profile Builder

Image format options

Image filename options

Image metadata options

Document metadata options

Engine Status Indicators

Using Monitor (Real-Time)

Using View

Appendix A: Image Processing Operations