Image analysis – Color image processing
Reexamination Certificate
2002-10-25
2004-08-10
Mariam, Daniel (Department: 2621)
Image analysis
Color image processing
C382S165000, C348S029000, C348S034000
Reexamination Certificate
active
06775406
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to assisting in the analysis of an image and, in particular, to facilitating the identification of a pattern in an image. More particularly, the invention relates to the finding of an anomaly in a black-and-white image, such as an x-ray image (e.g., medical or dental x-ray images).
2. Related Art
The following are explanations of words, phrases, abbreviations and concepts used throughout this description:
1. Grayscale: A black-and-white image is displayed in shades of gray varying from no gray (white) to fully saturated gray (black). The various shades of gray in between (intensity of gray) are based on a mix of black and white. A value in this intensity range can be represented in bits.
2. Bits: A set of binary representations (each bit is a single binary representation). For example, a grayscale value can be represented by a series of bits. The number of bits establishes the number of gradations between black and white in the grayscale (i.e., the number of possible grayscale values). The number of possible grayscale values is equal to 2
n
, where n is the number of bits. For example, a 2 bit grayscale can represent 4 shades of gray (including black and white), while an 8 bit grayscale can represent 256 grades of gray (again, including black and white).
3. Pixel: A very small dot in an image (e.g., an image displayed on a computer screen or a photograph) or a virtual dot within a computers' memory representing a very small dot in an image that can be displayed on a computer screen. The density of pixels in an image represents the image's resolution. Many black-and-white photographic images have a resolution of over 5000 pixels to the inch. A black-and-white image can be scanned into a computer's memory at, for example, a resolution of about 1200 to 2400 pixels to the inch. Many computer display monitors can only display about 100 to 200 pixels to the inch;
however, most computers' memory can still retain a scanned image resolution of 1200 to 2400 pixels to the inch. In a black-and-white image, each pixel can have any of a variety of shades of gray as represented by a series of bits.
4. DPI: Dots per inch, or pixels per inch.
5. Color roll: When a computer display monitor can only display 200 DPI and the image has 1200 DPI, the image includes six times the information that can be displayed at any one location. As described in more detail below, when the image information is represented in color (i.e., the image is colorized), display of all of the image information can be achieved by displaying over time a series of versions of the image in which the color displayed in each version of the image is alternated.
6. Anomaly: Most things in nature have smooth flowing lines and curves. A flaw in this smoothness can be referred to as an anomaly. A good example of an anomaly is a cut or scar on a person's forearm that interrupts the smooth surface of the forearm. As discussed in more detail below, the invention can advantageously be used to identify anomalies in an image (and, in particular, in an image of a part of the human anatomy).
7. Scanned image: Using a piece of equipment called a scanner, a black-and-white (or color) image on paper (or other similar medium) can be transformed from the “picture” to binary data that is usable by a computer. The computer representation of the picture is called a scanned image.
The expertise needed to locate a small (e.g., at the millimeter level) anomaly in an image takes a great deal of training and talent. In particular, discerning a small anomaly that is represented by less than 1% differential in grayscale can be extremely difficult. What has been needed is some computer assistance in performing that task, so that a person of lesser talent and/or experience can successfully screen an image, locate an anomaly, and pass on the findings to a specialist for further review.
FIG. 1
illustrates the pixels of a part of a black-and-white image. Each pixel is represented by a square box (160 pixels are illustrated in
FIG. 1
) and is colored by a shade of gray corresponding to a grayscale value for that pixel. It is assumed that the grayscale values for the image can range between 1 and 1000, 1 corresponding to white and 1000 corresponding to black. In
FIG. 1
, the intensity of the shade of gray for each pixel corresponds to the numerical value in the box representing that pixel. As can be seen, the lightest shade of gray in
FIG. 1
corresponds to the pixel having a grayscale value of
555
(in the upper left corner of
FIG. 1
) and the darkest shade of gray in
FIG. 1
corresponds to the pixel having a grayscale value of 571 (near the lower right corner of FIG.
1
). Thus, the grayscale intensity varies by a maximum of just 1.6% in the part of the black-and-white image illustrated in FIG.
1
. The human eye can only see about a 2.0% variation in grayscale intensity at best (and typically more like about 3.0%) and that only when the regions differing in grayscale intensity are side by side. While, as can be seen from the numerical values representing the grayscale intensities of the pixels in
FIG. 1
, there is a definite variation in grayscale intensity within the part of the black-and-white image illustrated in
FIG. 1
, that variation cannot be discerned by the human eye. Moreover, size of the pixels in
FIG. 1
is greatly exaggerated; pixels are typically much smaller, exacerbating the difficulty in perceiving variation in the grayscale intensity.
FIG. 2
illustrates the pixels of
FIG. 1
with each pixel colored with a color corresponding to the grayscale value for that pixel, i.e., the image is “colorized.” As illustrated in
FIG. 2
, three colors are used: red, green and blue. The three colors are repetitively assigned to grayscale values in order, i.e., as illustrated in
FIG. 2
, red is assigned to a grayscale value of
555
, green is assigned to a grayscale value of
556
, blue is assigned to a grayscale value of
557
, red is assigned to a grayscale value of
558
, green is assigned to a grayscale value of
559
, blue is assigned to a grayscale value of
560
, etc. In
FIG. 2
, all pixels are displayed at the same time with their corresponding color. While some patterning is discernible, this view of the image is still confusing since the same color is assigned to many different gray levels. A new approach to colorizing a black-and-white image to facilitate the identification of a pattern in the image is needed.
SUMMARY OF THE INVENTION
The invention facilitates the identification of a pattern in an image. In one embodiment of the invention, the identification of a pattern in an image can be facilitated by: 1) identifying a grayscale value for each of a multiplicity of pixels of the image (which can be all of the pixels of the image); 2) assigning one of a multiplicity of colors to each grayscale value; and 3) displaying a multiplicity of versions of the image in sequence at a specified rate, such that a) the display of each version of the image includes only pixels having a specified grayscale value or a grayscale value in a specified range of grayscale values that is smaller than the range of grayscale values for all pixels of the image, b) each pixel displayed in a version of the image is displayed in the color corresponding to the grayscale value of that pixel, and c) the specified grayscale value or range of grayscale values for which pixels are displayed in each successive displayed version of the image is incremented or decremented from the specified grayscale value or range of grayscale values for which pixels are displayed in the immediately previous displayed version of the image by a specified grayscale value increment or decrement. The specified grayscale value increment or decrement can be, for example, 1. The invention can further be implemented so that at least one of the multiplicity of colors is assigned to more than one grayscale value. For example, the multiplicity of colors can be repetitively assigned in a specified
LandOfFree
Colorizing a black-and-white image to facilitate the... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Colorizing a black-and-white image to facilitate the..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Colorizing a black-and-white image to facilitate the... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3294351