Manufacturing imperfections of photoconductor (PC) drums in electrophotographic (EP) printers cause low-
frequency artifacts that could produce objectionable non-uniformities in the final printouts. In this paper, we
propose a technique to detect and quantify PC artifacts. Furthermore, we spatially model the PC drum surface
for dynamic compensation of drum artifacts. After scanning printed pages of flat field areas, we apply a wavelet-
based filtering technique to the scanned images to isolate the PC-related artifacts from other printing artifacts,
based on the frequency, range, and direction of the PC defects. Prior knowledge of the PC circumference
determines the printed area at each revolution of the drum for separate analysis. Applied to the filtered images,
the expectation maximization (EM) algorithm models the PC defects as a mixture of Gaussians. We use the
estimated parameters of the Gaussians to measure the severity of the defect. In addition, a 2-D polynomial fitting
approach characterizes the spatial artifacts of the drum, by analyzing multiple revolutions of printed output.
The experimental results show a high correlation of the modeled artifacts from different revolutions of a drum.
This allows for generating a defect-compensating profile of the defective drum.
Print mottle is one of several attributes described in ISO/IEC DTS 24790, a draft technical specification for the measurement of image quality for monochrome printed output. It defines mottle as aperiodic fluctuations of lightness less than about 0.4 cycles per millimeter, a definition inherited from the latest official standard on printed image quality, ISO/IEC 13660. In a previous publication, we introduced a modification to the ISO/IEC 13660 mottle measurement algorithm that includes a band-pass, wavelet-based, filtering step to limit the contribution of high-frequency fluctuations including those introduced by print grain artifacts. This modification has improved the algorithm’s correlation with the subjective evaluation of experts who rated the severity of printed mottle artifacts. Seeking to improve upon the mottle algorithm in ISO/IEC 13660, the ISO 24790 committee evaluated several mottle metrics. This led to the selection of the above wavelet-based approach as the top candidate algorithm for inclusion in a future ISO/IEC standard. Recent experimental results from the ISO committee showed higher correlation between the wavelet-based approach and the subjective evaluation conducted by the ISO committee members based upon 25 samples covering a variety of printed mottle artifacts. In addition, we introduce an alternative approach for measuring mottle defects based on spatial frequency analysis of wavelet- filtered images. Our goal is to establish a link between the spatial-based mottle (ISO/IEC DTS 24790) approach and its equivalent frequency-based one in light of Parseval’s theorem. Our experimental results showed a high correlation between the spatial and frequency based approaches.
When evaluating printer resolution, addressability is a key consideration. Addressability defines the maximum number of spots or samples within a given distance, independent of the size of the spots when printed. Effective addressability is the addressability demonstrated by the final, printed output. It is the minimum displacement possible between the centers of printed objects. In this paper, we present a measurement procedure for effective addressability that offers an automated way to experimentally determine the addressability of the printed output. It requires printing, scanning, and measuring a test target. The effective addressability test target contains two types of elements, repeated to fill the page: fiducial lines and line segments. The fiducial lines serve as a relative reference for the incremental displacements of the individual line segments, providing a way to tolerate larger-scale physical distortions in the printer. An ordinary reflection scanner captures the printed test target. By rotating the page on the scanner, it is possible to measure effective addressability well beyond the scanner’s sampling resolution. The measurement algorithm computes the distribution of incremental displacements, forming either a unimodal or bimodal histogram. In the latter case, the mean of the second (non-zero) peak indicates the effective addressability. In the former case, the printer successfully rendered the target’s resolution, requiring another iteration of the procedure after increasing the resolution of the test target. The algorithm automatically estimates whether the histogram is unimodal or bimodal and computes parameters describing the quality of the measured histogram. Several experiments have refined the test target and measurement procedure, including two round-robin evaluations by the ISO WG4 committee. Results include an analysis of approximately 150 printed samples. The effective addressability attribute and measurement procedure are included in ISO/IEC TS 29112, a technical specification that describes the objective measurement of printer resolution for monochrome electrophotographic printers.
Grain is one of several attributes described in ISO/IEC TS 24790, a technical specification for the measurement of
image quality for monochrome printed output. It defines grain as aperiodic fluctuations of lightness greater than
0.4 cycles per millimeter, a definition inherited from the latest official standard on printed image quality, ISO/IEC
13660. Since this definition places no bounds on the upper frequency range, higher-frequency fluctuations (such
as those from the printer’s halftone pattern) could contribute significantly to the measurement of grain artifacts.
In a previous publication, we introduced a modification to the ISO/IEC 13660 grain measurement algorithm
that includes a band-pass, wavelet-based, filtering step to limit the contribution of high-frequency fluctuations.
This modification improves the algorithm’s correlation with the subjective evaluation of experts who rated the
severity of printed grain artifacts.
Seeking to improve upon the grain algorithm in ISO/IEC 13660, the ISO/IEC TS 24790 committee evaluated
several graininess metrics. This led to the selection of the above wavelet-based approach as the top candidate
algorithm for inclusion in a future ISO/IEC standard. Our recent experimental results showed r2 correlation
of 0.9278 between the wavelet-based approach and the subjective evaluation conducted by the ISO committee
members based upon 26 samples covering a variety of printed grain artifacts. On the other hand, our experiments
on the same data set showed much lower correlation (r2 = 0.3555) between the ISO/IEC 13660 approach and
the same subjective evaluation of the ISO committee members.
In addition, we introduce an alternative approach for measuring grain defects based on spatial frequency analysis
of wavelet-filtered images. Our goal is to establish a link between the spatial-based grain (ISO/IEC TS 24790)
approach and its equivalent frequency-based one in light of Parseval’s theorem. Our experimental results showed
r2 correlation near 0.99 between the spatial and frequency-based approaches.
Improper design of color halftone screens may create visually objectionable moire patterns in the final prints due
to the interaction between the halftone screens of the color primaries. The prediction of such interactions from the
screens' bitmaps helps to identify and avoid problematic patterns, reducing the time required to design effective
color halftone screens. In this paper, we detect the moire patterns by examining the spatial frequency spectra of
the superimposed screens. We study different windowing techniques including Hann, Hamming, and Blackman,
to better estimate the moire strength, frequency and orientation. The window-based spectral estimation has
the advantage of reducing the effect of spectral leakage associated with the non-windowed discrete signals. Two
methods are used to verify the detected moire from the bitmaps. First, we analyze scans of the printed halftones,
using the same technique that we applied to the bitmaps. Second, we independently inspect the printed halftones
visually. Our experiments show promising results by detecting the moire patterns from both the bitmap images
as well as the scans of the actual prints verified by visual inspection.
Several measurable image quality attributes contribute to the perceived resolution of a printing system. These
contributing attributes include addressability, sharpness, raggedness, spot size, and detail rendition capability. This
paper summarizes the development of evaluation methods that will become the basis of ISO 29112, a standard for the
objective measurement of monochrome printer resolution.
In this paper, we propose new techniques for detecting and quantifying print defects. In our previous work, we
introduced a scanner-based print quality system to characterize directional print defects, such as banding, jitter,
and streaking. We extend our previous print quality work two ways. First, we introduce techniques for detecting
2-D isotropic, mottled print defects such as grain and mottle. Wavelet pre-filtering is used to limit the defect's
size or frequency range. Then we analyze the L* variation in the wavelet-processed images. The methods used
to quantify grain and mottle are similar to ISO/IEC 13660 techniques. The second part of this paper provides
techniques for detecting and quantifying low frequency directional defects, which we call left-to-right and
top-to-bottom L* variation. Since these defects extend less than two cycles across the page, and probably less than
a complete cycle, we fit a 4th-degree polynomial to the defect profile. To measure the strength of the defect,
we use variational analysis of the fitted polynomial. Experimental results on 10 printers and 100 print samples
showed an average correlation for isotropic defects of 0.85 between the proposed measures and experts' visual
evaluation, and 0.97 for low frequency defects.
In this paper we present a unified framework for physical print quality. This framework includes a design for
a testbed, testing methodologies and quality measures of physical print characteristics. An automatic belt-fed
flatbed scanning system is calibrated to acquire L* data for a wide range of flat field imagery. Testing
methodologies based on wavelet pre-processing and spectral/statistical analysis are designed.
We apply the proposed framework to three common printing artifacts: banding, jitter, and streaking. Since
these artifacts are directional, wavelet based approaches are used to extract one artifact at a time and filter
out other artifacts. Banding is characterized as a medium-to-low frequency, vertical periodic variation down
the page. The same definition is applied to the jitter artifact, except that the jitter signal is characterized as
a high-frequency signal above the banding frequency range. However, streaking is characterized as a horizontal
aperiodic variation in the high-to-medium frequency range.
Wavelets at different levels are applied to the input images in different directions to extract each artifact within
specified frequency bands. Following wavelet reconstruction, images are converted into 1-D signals describing
the artifact under concern. Accurate spectral analysis using a DFT with Blackman-Harris windowing technique
is used to extract the power (strength) of periodic signals (banding and jitter). Since streaking is an aperiodic
signal, a statistical measure is used to quantify the streaking strength.
Experiments on 100 print samples scanned at 600 dpi from 10 different printers show high correlation (75%
to 88%) between the ranking of these samples by the proposed metrologies and experts' visual ranking.
Proc. SPIE. 6064, Image Processing: Algorithms and Systems, Neural Networks, and Machine Learning
KEYWORDS: Image processing algorithms and systems, Signal to noise ratio, Detection and tracking algorithms, Image segmentation, Scanners, Linear filtering, Feature extraction, Halftones, Prototyping, Fuzzy logic
In this paper, we introduce a new system to segment and label document images into text, halftoned images, and background using a modified fuzzy c-means (FCM) algorithm. Each pixel is assigned a feature vector, extracted from edge information and gray level distribution. The feature pattern is then assigned to a specific region using the modified fuzzy c-means approach. In the process of minimizing the new objective function, the neighborhood effect acts as a regularizer and biases the solution towards piecewise-homogeneous labelings. Such a regularization is useful in segmenting scans corrupted by scanner noise.
In this paper, we present a new system to segment and label the contents of scanned documents as either text or image, using a modified fuzzy c-means (FCM) algorithm. Each pixel is assigned a feature pattern extracted from the gray level distribution and computed at different scales. The invariant feature pattern is then assigned to a specific region using fuzzy logic. Our algorithm is formulated by modifying the objective function of the standard FCM algorithm to allow the labeling of a pixel to be influenced by the labels in its immediate neighborhood. The neighborhood effect acts as a regularizer and biases the solution towards piecewise-homogeneous labelings. Such a regularization is useful in segmenting scans corrupted by scanner noise.
Image resizing is an important operation that is used extensively in document processing to magnify or reduce images. Standard approaches fit the original data with a continuous model and then resample this 2D function on a few sampling grid. These interpolation methods, however, apply an interpolation function indiscriminately to the whole image. The resulting document image suffers from objectionable moire patterns, edge blurring and aliasing. Therefore, image documents must often be segmented before other document processing techniques, such as filtering, resizing, and compression can be applied. In this paper, we present a new system to segment and label document images into text, halftone images, and background using feature extraction and unsupervised clustering. Once the segmentation is performed, a specific enhancement or interpolation kernel can be applied to each document component. In this paper, we demonstrate the power of our approach to segment document images into text, halftone, and background. The proposed filtering and interpolation method results in a noticeable improvement in the enhanced and resized image.
Green noise is the mid-frequency component of white noise and has been shown to have visually pleasing attributes when applied to digital halftoning. Unlike blue noise dither patterns, which are composed exclusively of isolated pixels, green noise dither patterns are composed of pixel-clusters making them less susceptible to image degradation from non- ideal printing artifacts such as dot-loss. Clearly, these patterns reduce the spatial variation in tone produced by electrophotographic printers when printing a constant shade of gray, but to date, no study has been presented showing the amount of reduction. In this paper, we address this problem by studying the effects of changing the average cluster size in a green noise dither pattern, measuring the resulting spatial variations for a Lexmark Optra laser printer in 1200 dpi mode. The print quality is evaluated in terms of the visibility of printer mechanism noise and the average change in tone across the printed page.
In this paper, we present a new system to segment and label document images into text, halftone images, and background using feature extraction and unsupervised clustering. Each pixel is assigned a feature pattern consisting of a scaled family of differential geometrical invariant features and texture features extracted from the cooccurence matrix. The invariant feature pattern is then assigned to a specific region using a two-stage neural network system. The first stage is a self-organizing principal components analysis (SOPCA) network that is used to project the feature vector onto its leading principal axes found by using principal components analysis. Using the SOPCA algorithm, we can train the SOPCA network to project our feature vector orthogonally onto the subspace spanned by the eigenvectors belonging to the largest eigenvalues. By doing that we ensure that the vector is represented by a reduced number of effective features. The next step is to cluster the output of the SOPCA network into different regions. This is accomplished using a self-organizing feature-map (SOFM) network. In this paper, we demonstrate the power of the SOPCA-SOFM approach to segment document images into text, halftone, and background.
The use of model-based approaches in the detection and identification of man-made objects is encountered frequently in the literature today. These approaches generally depend on very high resolution imagery, and must first be cued to an approximate location of these objects. The approaches are also highly computationally intensive and, therefore, cannot be relied upon to perform broad sweeps in either a real time scenario or during training stages. It is therefore necessary to preprocess imagery with techniques that are less expensive in computer processing and more general in their approach to detection. With these requirements in mind we examine a modified form of the chord transform in the detection of man-made objects in an image. In addition, since the chord transform is an O(N4) algorithm, we explore the possibility of a parallel implementation of this approach in a SIMD architecture. The resulting mechanism is capable of quickly identifying straight lines, right angles, parallel lines, and arcs in an image. These primitives are indicative of man-made objects in an image.