Multimedia Graphics and Image Data Representations
Graphics/Image Data Types
- Proliferation of file formats in multimedia.
1-bit Images
- Each pixel is a single bit (0 or 1), called binary image.
- Also called 1-bit monochrome image (no color).
- Example: "Lena" image.
8-bit Gray-level Images
- Each pixel has a gray-value between 0 and 255 (single byte).
- Bitmap: 2D array of pixel values.
- Image resolution: Number of pixels (higher is better).
- High resolution: 1,600×1,200.
- Low resolution: 640×480.
- Frame buffer: Hardware to store bitmap (video card).
- 8-bit image as bit-planes: Each plane is a 1-bit representation at increasing levels.
- Pixel stored as a byte (0 to 255).
- 640×480 grayscale image: 300 kB.
- Dithering: Trades intensity resolution for spatial resolution on 1-bit printers.
Dithering
- Dithering: Calculates dot patterns for printing on 1-bit printers.
- Replaces pixel value by larger pattern (e.g., 2×2 or 4×4).
- Example dither matrix: [0amp;2 3amp;1]
- Ordered dither: Turns on printer bit if intensity is greater than matrix element at pixel position.
- Algorithm:
- Iterate through image pixels, compare intensity to dither matrix.
- O(x,y)=1 if I(x, y) > D(i, j), else O(x,y)=0
Image Data Types
- Common data types: 24-bit color and 8-bit color.
- Formats can be platform-specific or cross-platform.
- Image formats use compression (lossless or lossy).
24-bit Color Images
- Each pixel is three bytes (RGB).
- 256×256×256=16,777,216 possible colors.
- 640×480 24-bit image: 921.6 kB.
- Some images use 32-bit (extra byte for alpha value/transparency).
8-bit Color Images
- Uses a lookup table to store color information.
- Bytes are indices into a table with 3-byte (color) values.
- 640×480 8-bit color image: 300 kB.
Color Look-up Tables (LUTs)
- Stores index/code value for each pixel.
How to devise a color look-up table
- Divide the RGB cube into equal slices.
- Scale R and G to 3-bit range (0..7), B to 2-bit range (0..3).
- Median-cut algorithm: Sort R values, label with "0" or "1" based on median.
GIF
- Limited to 8-bit (256) color images.
- Supports interlacing (4-pass display).
- Two flavors: GIF87a (original), GIF89a (animation support).
JPEG
- Most important standard for image compression.
- Exploits limitations of human vision.
- User can set quality level/compression ratio.
PNG
- Portable Network Graphics; supersedes GIF.
- Up to 48 bits of color.
- Gamma-correction, alpha-channel.
- Progressive display (2D fashion, 7 passes through 8×8 blocks).
TIFF
- Tagged Image File Format.
- Supports additional information via “tags”.
- Can store 1-bit, grayscale, 8-bit, 24-bit RGB images.
- Originally lossless, now supports JPEG compression.
EXIF
- Exchange Image File: for digital cameras.
- Uses JPEG format.
- Tags for camera/picture-taking conditions.
Graphics Animation Files
- Stores graphics animations (series of drawings).
- Less demanding than video files.
- Examples: FLC, FLI, GL, animated GIF89.
PS and PDF
- Postscript: Language for typesetting (vector-based).
- Includes text and vector graphics.
- PDF: Portable Document Format (includes LZW compression).
- WMF (Windows Metafile): Vector format for Windows.
- BMP: Major system standard for Windows.
- PAINT and PICT (Macintosh).
- PPM : graphics format for the X Window system