SEMINAR

ABSTRACT

Monitors and printers for displaying color images are very popular today. Most of these devices are limited to working with an undetermined palette of 256 colors. Many images that need to be displayed are composed of far more colors than these devices can handle. Therefore, it is imperative to have a method for reducing the number of colors contained in an image in order to display that image on a device with limited color capabilities.

The application of an algorithm for reducing the number of colors used in an image is called color quantization. The ultimate result of color quantization is the mapping of each color in the original image to a new color in a limited or representative color map, while maintaining the appearance of the original image. The analytical quantization error measurement can be used to determine how well the quantized image resembles the original. This measurement can also be used to compare different quantization algorithms.

The area of quantization in computer graphics is derived from the work in the areas of digital image processing and communications. In these areas, quantization is dealt with in both the gray-scale and color domains. Much of the early work on color quantization can be traced back to the research and development of the color television. The main concern of quantization in this area is the reduction of the bandwidth for broadcasting or transmitting color video signals.

Color quantization research in communications is once again very popular with the current research focused on ''High-Definition'' television. The area of digital image processing is focused more on gray-scale quantization including parts of the invisible spectrum such as ultra-violet intensities. Quantization is also used in other areas of computer graphics, including data compression, edge detection and volume rendering.

The image produced by the quantization process can contain artifacts resulting from the lower intensity resolution. The most common artifact is that of contouring. Contouring occurs when the color intensities in two adjacent regions differ significantly without a gradual change from one intensity to the next.

Artifacts in an image degrade the quality of that image. One approach to reducing these artifacts is to smooth the image by blending neighboring pixels. A common approach is to use the error-diffusion dithering technique in the quantization process to create the illusion of color tones where none actually exists. Reducing these artifacts is to smooth the image by blending neighboring pixels. A common approach is to use the error-diffusion dithering technique in the quantization process to create the illusion of color tones where none actually exists.

WHERE: TEC 340

WHEN(day): Friday, March 6th, 1998

WHEN(time): 2:00 PM

EVERYBODY IS INVITED