Computer graphics processing and selective visual display system – Computer graphics processing – Attributes
Reexamination Certificate
2001-09-28
2004-05-04
Bella, Matthew C. (Department: 2676)
Computer graphics processing and selective visual display system
Computer graphics processing
Attributes
C345S593000, C345S600000, C345S605000
Reexamination Certificate
active
06731299
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a method for converting N-bit image data of a digital image into M-bit image data, in which N−M=K, K>0. More particularly, the present method is applied to a dithering process for computer graphics or image processing to solve the image deviation problem.
BACKGROUND OF THE INVENTION
Converting images from high precision into low precision is commonly used in computer graphics and image processing. As known, a processor is capable of processing an image with high precision data, a storage or display unit cooperating therewith, however, may be limited in low precision. This is because an image processor is designed to support various display standards, e.g., 24-bit true color format, and 16-bit high color format. Thus, it is often desirable to convert 24-bit data generated by the image processor to 16-bit data suitable for storing and/or displaying.
The 24-bit true color format takes 8-bit for representing the red color, 8-bit for representing the green color, and 8-bit for representing the blue color. On the other hand, the 16-bit high color format takes 5-bit, 6-bit and 5-bit for representing the red, green and blue colors, correspondingly. Therefore, when a pixel with true color format is converted into that with high color format, it needs to perform 8-to-5 and 8-to-6 conversions, as shown in FIG.
1
.
Conventionally, truncation and rounding methods are used for such conversions. Simple truncation and rounding, however, are undesirable due to Mach banding effects that are introduced into the final image. Regarding the Mach banding effect, an intensity gap occurs at any edge where there is discontinuity in magnitude or slope of intensity, and it is exaggerated by the human eye. To avoid the Mach banding effect, a dithering process is developed to exploit spatial integration in the eye. In the dithering operation, a dither value from a dither matrix is combined with a color value prior to truncation. The dither value varies with pixel location in the image. FIGS.
2
A~
2
C shows three examples of dither matrices or dither tables for 2-bit, 3-bit and 4-bit down conversions from m bits, respectively, in which i and j indicates the x- and y-coordinates of the pixel, and i=(x mod m), j=(y mod m).
By the use of the dither tables, random values are added to pixel intensities to break up contours, which are referred to as dither noise. Various algorithms have been used to generate the random distributions of the dither tables. By adding the dither noise over an entire picture, the intensity boundaries can be softened.
Please refer to
FIG. 3
which is a schematic functional block diagram of a conventional dithering process for 8-to-5 conversion of a pixel, as applied to the red color conversion from a true color format to a high color format. First of all, the coordinate data of the pixel, i.e. x and y, are entered the functional block
31
to correspond to a value in the dither table D
3
of
FIG. 2B
to determine a threshold value T. In the functional block
32
, the threshold value T is compared with a value of the least significant 3 bits R[2:0] of the 8-bit color value R[7:0] of the pixel to realize a carry value. When the 3-bit color value R[2:0] is greater than the threshold value T, the carry value is 1, and the most significant 5 bits R[7:3] of the 8-bit color value of the pixel is added thereto the carry value 1 in the functional block
33
to form a temporary value denoted as Tmp[5:0]. On the contrary, when the 3-bit color value R[2:0] is less than or equal to the threshold value T, the most significant 5 bits R[7:3] of the 8-bit color value of the pixel is added thereto the carry value 0. That is, the temporary value Tmp[5:0] is equal to the unchanged most significant 5 bits R[7:3]. The temporary value Tmp[5:0] is further processed by a clamp operation in the functional block
34
to convert the temporary value Tmp[5:0] into the desired 5-bit color value R′[4:0] of the pixel. For the clamp operation, if the value of Tmp[5:0] is greater than 31, the dithered value R′[4:0] is set to be 31. Otherwise, the dithered value R′[4:0] is equal to the value of Tmp[5:0]. It is to be noted that Tmp[5:0] instead of Tmp[4:0] is used herein because the temporary value may overflow after the addition of the carry value.
FIG. 4
shows how the conventional dithering process similar to
FIG. 3
performs 8-to-6 conversion of a pixel, as applied to the green color conversion from a true color format to a high color format. First of all, the coordinate data of the pixel, i.e. x and y, are entered the funvtional block
41
to correspond to a value in the dither table D
2
of
FIG. 2A
to determine a threshold value T. The threshold value T is compared with a value of the least significant 2 bits G[1:0] of the 8-bit color value G[7:0] of the pixel in the functional block
42
. When the 2-bit color value G[1:0] is greater than the threshold value T, the most significant 6 bits G[7:2] of the 8-bit color value of the pixel is added thereto a carry value of 1 in the functional block
43
to form a temporary value denoted as Tmp[6:0]. On the contrary, when the 2-bit color value G[1:0] is less than or equal to the threshold value T, the most significant 6 bits G[7:2] of the 8-bit color value of the pixel is added thereto a carry value of 0. That is, the temporary value Tmp[6:0] is equal to the unchanged most significant 6 bits G[7:2]. The temporary value Tmp[6:0] is further processed by a clamp operation in the functional block
44
to convert the temporary value Tmp[6:0] into the desired 6-bit color value G′[5:0] of the pixel. For the clamp operation, if the value of Tmp[6:0] is greater than 63, the dithered value G′[5:0] is set to be 63. Otherwise, the dithered value G′[5:0] is equal to the value of Tmp[6:0]. It is to be noted that Tmp[6:0] instead of Tmp[5:0] is used herein because the temporary value may overflow after the addition of the carry value.
On the contrary to the conversion illustrated with reference to
FIG. 3
or
4
, an inverse dithering operation is typically used for converting the image from the high color format to the true color format. A typical inverse dithering operation is to append several the most significant bits as illustrated in the functional block
51
or
52
of
FIG. 5A
or
5
B. In
FIG. 5A
, the recovery value R″[7:0] is obtained by using the 5-bit color value R′[4:0] as the most significant 5 bits R″[7:3], and complementing the color value R″[7:0] by adding the most significant 3 bits R′[4:2] as the least significant 3 bits R″[2:0]. Likewise, in
FIG. 5B
, the recovery value G″[7:0] is obtained by using the 6-bit color value G′[5:0] as the most significant 6 bits G″[7:2], and complementing the color value G″[7:0] by adding the most significant 2 bits G′[5:4] as the least significant 2 bits G″[1:0].
The dithering method mentioned above achieves better visual quality than directly truncation for precision conversion. However, for computer graphics applications, a dithered image needs to be inversely dithered for further processing and then dithered again. A conventional inversely dithering method is to append the most significant bits into the dithered value for low to high precision conversion. Unfortunately, after repeating the dithering and inversely dithering processes several times, some points of the resulting image appear to become too red, too green or too blue. That is, a color deviation problem
Lu Chung-Yen
Wang Hai-Wei
Bella Matthew C.
Chen Po-Wei
Lowe Hauptman & Gilman & Berner LLP
Silicon Integrated Systems Corp.
LandOfFree
Apparatus and method for dithering in image processing and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for dithering in image processing and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for dithering in image processing and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3203946