Next |
Prev |
Top
|
JOS Index |
JOS Pubs |
JOS Home |
Search
- Compresses individual images (no motion prediction as in MPEG)
- Baseline JPEG quantizes 2D DCT of
block of pixels
- Specialized, optimized FFT-like DCT transforms used
- Colors processed separately
- DCT blocks ordered in fixed ``raster'' pattern
- DCT approximates the Karhunen-Loeve transform (equal in the limit as transform size
)
- Compression ratio variable
- Progressive coding supported
- Low-frequency DCT coefficients sent first
- Higher frequency DCT coefficients sent later
- Hierarchical (``pyramidal'') resolution coding supported (HF coding differential wrt LF)
- Lossless predictive coding also supported (no DCT)
- ``Blocking'' artifacts possible due to non-overlapping DCT blocks
- Uses entropy coding (Huffman)
Next |
Prev |
Top
|
JOS Index |
JOS Pubs |
JOS Home |
Search
Download Compression.pdf
Download Compression_2up.pdf
Download Compression_4up.pdf
[Comment on this page via email]