Radiant energy – Photocells; circuits and apparatus – Optical or pre-photocell system
Reexamination Certificate
1999-10-26
2002-04-23
Lee, John R. (Department: 2878)
Radiant energy
Photocells; circuits and apparatus
Optical or pre-photocell system
C250S235000, C358S451000
Reexamination Certificate
active
06376834
ABSTRACT:
BACKGROUND OF THE INVENTION
Scanners are devices that view an image and then represent it as if it were a number of adjacent lines of consecutive pixel positions along a line. The optical content at each pixel location is represented by a digital value (say, four or five bits), and the scanner's output is often organized serially, as consecutive digital values delimited or grouped into collections representing the scanned lines. An important parameter in this process is the spatial density represented by consecutive digital values within a line, often expressed in dots per inch, or DPI. A similar parameter is the distance between adjacent lines, which may be termed lines per inch, or LPI. For most scanners the DPI and LPI are fixed during any given scanning operation; for certain scanners the LPI may vary in complicated ways if during scanning the LPI axis is allowed to depart from perpendicularity with the DPI axis. Sometimes firmware associated with preparing the gray scale data for use by another system re-formats the scanned results to appear as though they had been scanned at a lower resolution than actually performed by the hardware. In any event, the scanner system usually outputs pixel data that is expected to already conform, or that has been adjusted to conform, to a regular grid of pixel locations. For a variety of different light sensitive sensors, this scanning process is capable of very good performance under favorable conditions.
Leaving aside a host of performance affecting issues that stem from the nature of the light sensitive sensor, one very important issue attaches to the nature of the original image being scanned. As is well known, many images that begin as photographs are reproduced as half-tones. Under magnification the half-toning appears as a regularly spaced matrix of dots having different sizes. Upon such examination (say, with a magnifying glass) it is seen that dark dots are really no darker than light ones, but they are instead bigger. In fact, adjacent dark dots might be so large as to blend together, so that there is an absence of intervening white space. Really light dots are sufficiently small (to the point of absence, even) that they have lots of intervening white space.
Half-toning is often produced by the introduction of a screen in the optical path. The holes in the screen are apertures that produce the dots of varying size. The pitch of the screen (i.e., the number of holes per inch) determines (absent other enlargement or reduction) the spacing for the matrix of dots in the half-tone reproduction, which might be the printed image of a photograph in a newspaper or a magazine. The pitch of the screen is generally uniform in both directions. The screen used for newspaper photographs might be as coarse as eighty five lines per inch, or less, while that for an expensive magazine might be one hundred fifty lines per inch, or more.
Half-toning solves some otherwise very nasty problems in the reproduction of photographs, and is ubiquitous. Printers for a computer often imitate the half-tone process to increase the fidelity with which a regular office copier can reproduce the printer's output of an image composed of elements besides the easy-to-copy solid lines or “type” (e.g., the 100% solid go characters of a font).
Consider a laser printer. It might have an internal micro-dot resolution of 600 or 1200 DPI, which is used to create groups of larger dots of varying size. There are various known ways to take gray scale data for pixels at, say, one hundred dots per inch, and create from the micro-dots an image of larger variable size dots at, say, one hundred dots per inch. One of these process is known as “ordered dithering”, and uses the “fine positions” (X and Y positions mod
n
) of an incoming pixel to address an n by n matrix of varying threshold values. If the gray scale value of the incoming pixel is at or above the threshold a corresponding cluster of micro-dots is produced. For fields of truly all white and truly solid black, this process produces true absence of micro-dots and solid micro-dots, respectively. For the levels of gray between those extremes there will be clusters of micro-dots of variable size. It is as if the printer had its own internal screen, although in most cases the screening agent is the software driver, and not the printer itself
Moire patterns are the familiar regular pattern of dark bands or spots distributed across an image containing two (or more) underlying quantization regimes, such as having been screened by two different screens of different pitch. The relative orientation of the two regimes (screens) has an influence on the shape of the Moire pattern produced. When an image is viewed through two misaligned or different screens the Moire pattern relies on a regular disturbance in image intensity to reveal itself. That is, if two cells of the screens are in good alignment the “light” for the corresponding subject matter “gets through,”, as if there were only one screen; any other alignment increasingly interferes with image fidelity during transmission of the image. Now, the case that will interest us the most concerns the viewing (say, by printing) of an original unscreened image through a second screen (in the printer) after it has already been half-toned with a first screen. We can dispense with the first screen if we take into account that what impinges on the second screen is a pattern of variable size dots, separated by, or on a field of, white space. As before, the dots in a region might be so large as to blot out the white space, or the dots might be so small that the region is entirely (or nearly) white space.
Thus, if the second screen aligns completely, or temporarily aligns well in a particular region, with the half-toned dots of the image, then within the region of good alignment there are no artifacts to disturb the transmission of the image; the dots of the image produce corresponding clusters of printed micro-dots. Now, suppose that the second screen is significantly or totally misaligned with the dots of the image. If the second screen is the same screen as the first, but simply translated out of maximum alignment (but not rotated), then the entire image would appear uniformly dim by an amount corresponding to the lack of alignment, as the apparent darkness or effective size of all the dots is uniformly reduced. If the second screen has a different pitch or is rotated, then there will be periodically occurring instances of good alignment interspersed with bad alignment. In regions of poor alignment the image either: (a) begins to wash out, as consecutive smallish dots fall in the spaces between consecutive apertures of the second screen (regularly fail to meet the threshold of the ordered dither); or (b) goes solid, as consecutive largish dots bridge across consecutive apertures of the second screen (regularly exceed the threshold of the ordered dither). In case (a) small dots are removed (and medium size dot made smaller), while in case (b) medium size dots are made into larger dots. These disturbances to dot size are the visible Moire pattern.
The scanning and subsequent printing of previously reproduced images is becoming more and more prevalent. Unfortunately, this can be accompanied by two undesirable outcomes.
Simply scanning a half-toned image can cause detectable Moire pattern artifacts in the data of the gray scale values for the scanned result. They emerge along the DPI direction at a first spatial frequency that is the difference between the half-toning screen's effective pitch in that direction and the DPI of the scanner. They also emerge along the LPI direction at a second spatial frequency that is the difference between the half-toning screen's effective pitch in that direction and the LPI of the scanner. In other words, the resolution of the scanner's hardware acts as if it were a screen. Upon reflection it will be agreed it would be fair to call these “scanner induced” Moire patterns.
More noticeable than “scanner induced” Moire patterns is what happ
Dalton Dan L
Hastings Brian L
Hewlett--Packard Company
Lee John R.
LandOfFree
Moire interference pattern reduction for scanning of... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Moire interference pattern reduction for scanning of..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Moire interference pattern reduction for scanning of... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2868312