Image analysis – Image enhancement or restoration – Artifact removal or suppression
Reexamination Certificate
1999-08-18
2003-11-11
Grant, II, Jerome (Department: 2624)
Image analysis
Image enhancement or restoration
Artifact removal or suppression
C358S003230, C358S003260, C358S001900
Reexamination Certificate
active
06647151
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
The present invention relates generally to the display and printing of digital images and, more particularly, to methods of coalescing a mosaic of separate digital image tiles, each tile presented in the form of a discrete bitmap, into a single bitmap for artifact-free image enhancement, and to methods of testing tiles to determine when coalescence is contraindicated.
2. Description of Related Art
Computer generation and processing of digital images is a vital component of present information technology. Computers have displayed information to the outside world in progressively more complex formats from alphanumeric, monochrome charts and graphs, and presently in full color images comparable in quality with the best photographs. Digital creation, manipulation and display of images is an expanding field of computer technology showing no signs of deceleration. Very few application programs are made commercially available that do not have capabilities for creating, manipulating or using high quality images for ease of program use, simplicity of communication with the user, or for entertainment, games or advertising. Thus, the engineer concerned with computer display and printing technologies must consider full color (typically 24 bit) and high resolution (typically 600 dpi) images.
The present invention relates to the interface between two digital imaging technologies; digital image enhancement and the display, printing and/or assembly of composite images from subimages referred to as “tiling.” “Tiling” is the process by which a single image is constructed from several (or very many) separate components or “tiles” that are assembled into the proper spatial relationship to construct the full image. Thus, the complete image is constructed as a “mosaic” of individual tiles. Some application programs may construct separate tiles and display (or print) each tile in the proper spatial relationship to form the complete image.
Many areas of technology are presented with the problem of assembling a single coherent image from numerous component subimages. One typical example relates to satellite photography in which an orbiting satellite acquires numerous sequential images of the scene below. These separate images (“tiles”) are to be assembled into a single composite image (“mosaic”). The process of assembling tiles into a mosaic is to be done in such a manner that it is not apparent to the viewer or user of the final mosaic image that it was assembled from tiles. The work of Burt et. al. describes ways of assembling mosaic images from tiles in U.S. Pat. Nos. 5,488,674 and 5,649,032. Schemes for matching common features and aligning adjacent tiles are described.
Other areas of technology also involve the assembly of mosaic images from tiles. For example, Adelson (U.S. Pat. No. 4,661,986) describes the assembly of a three dimensional image from a collection of two dimensional images of the same scene. Assembling a composite image from a series of electron micrographs is described in the work of Vogt and Trenkle (U.S. Pat. No. 5,796,861). These examples are intended to be illustrative only and not exhaust all areas of technology in which mosaic images are assembled from tiles.
However, these examples of assembling an image from tiles share a common feature in so far as it is known from the start that the images relate to the same scene. Aerial, satellite panoramic video or still photography typically depict variations of a single scene. That is, the user knows or presumes that the image tiles ought to be assembled into a single, composite mosaic image in which the separate tiles form a non-disjoint image. In addition, it is often the case that automated schemes of mosaic construction utilize the fact that consecutive images overlap, simplifying the search for common features for matching. Distinct images typically call for separate and distinct uses of the methods described by these prior works.
In the display or printing of computer generated images, the situation is more complex.
FIG. 1
depicts a typical page that a computer user would wish to print (often in color, unlike FIG.
1
). Application programs may create numerous tiles that need to be joined to form the separate images of giraffe (FIG.
1
B), hot rod (FIG.
1
C), etc. However, separate tiles forming components of distinct images (hot rod
1
C, and hockey player 1D, for example) should not be joined but remain distinct images as the user has placed them on the page as depicted in FIG.
1
. Therefore, any automated scheme for assembling tiles into a mosaic image must have the capability of joining tiles for a single image, and not joining tiles when separate images occur on a single page. This should be contrasted with previous tiling methods such as those noted above in which typically only a single image is under consideration, analogous to the image of FIG.
2
.
Digital data for creating pages or screen displays must include in some form the particular location at which each picture element, or “pixel,” is to be displayed and the color to be displayed at that particular pixel. However, modem display and printing technology is more complex than a simple rendering of pixel-by-pixel information onto the appropriate page or screen location. Color mapping, image enhancements and many other image processing procedures may be employed to render the colors on the printed page so that they appear to match, enhance contrast and/or resolution and provide myriad other image enhancements. It is imperative that image enhancements be carried out without introducing artifacts into the image. That is, while enhancing one image feature (such as contrast) artifacts (such as lines or bands) should not be created so as to detract from the overall image quality. Additionally, such image processing procedures must be rapidly performed so as not to reduce substantially the printing speed of each page or introduce annoying delays into screen displays. The interaction of tiling technologies with image enhancement technologies is the general field of the present invention.
It is convenient to consider two classes of image alteration or enhancement. A first class relates to a “local” transformation of the information in the image. That is, an image is altered pixel by pixel such that the alteration (transformation) applied to any pixel does not depend on the characteristics, alteration or transformation of any other pixel. Color mapping is a typical example of such a local image transformation in which the numerical color value of the screen display is altered to produce a printed pixel having a perceived color as close as possible to the color perceived by the viewer on the screen display. Users desire color on a printed page to be the same as that on the screen. This requires an analysis of the different color perception qualities of the images created by different devices and appropriate corrections. Such color correction is a typical example of local transformation of one image pixel into another image pixel in which the characteristics of other pixels are not considered in creating or applying the transformation. Many other local transformations may be employed as well.
In contrast to local image transformations are a variety of image enhancing procedures that depend on an analysis of several (or many) pixels for determining the characteristics and parameters of the subsequent image transformation. Such transformation depending on more than a single local pixel we denote as “non-local” or “global.” Spatially sensitive filters are another name often used to denote image enhancements making use of more than the local pixel to be transformed in determining the operation to be applied. Spatially sensitive filtering applies a transformation to each pixel of the image in which the transformation applied to any particular pixel is determined by the properties of several (or many) pixels in addition to the pixel being transformed. Spatially sensitive filtering may use the properties of the entire image or bit
Stallbaumer Philip
Yuan Feng
LandOfFree
Coalescence of device independent bitmaps for artifact... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Coalescence of device independent bitmaps for artifact..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Coalescence of device independent bitmaps for artifact... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3158819