A New Reference Design Development Environment for JPEG 2000 Applications - PowerPoint PPT Presentation

1 / 9
About This Presentation
Title:

A New Reference Design Development Environment for JPEG 2000 Applications

Description:

Scalable image files -- no decompression needed for reformatting ... It offers a fast host interface. Options for future expansion ... – PowerPoint PPT presentation

Number of Views:35
Avg rating:3.0/5.0
Slides: 10
Provided by: billf9
Category:

less

Transcript and Presenter's Notes

Title: A New Reference Design Development Environment for JPEG 2000 Applications


1
A New Reference Design Development Environment
for JPEG 2000 Applications
  • Bill FinchCAST, Inc.

Warren MillerAVNET Design Services
DesignCon 2003January 28,2003
2
The Need for A New Standard
  • Higher useful compression ratios to conserve
    bandwidth
  • True loss less compression
  • Faster processing of larger images
  • Error resilience
  • A single standard for all of the above

3
JPEG 2000 Advantages
  • Better image quality at the same file size
  • 25-35 smaller file sizes at comparable image
    quality
  • Good image quality even at very high compression
    ratios, over 801
  • Low complexity option for devices with limited
    resources such as cell phones
  • Scalable image files -- no decompression needed
    for reformatting
  • Progressive rendering and transmission through a
    layered image file structure

4
Block Diagram of the JPEG 2000 Core
5
Considerations in Application Design
  • Image size (s) to be processed
  • Image composition (color, mono, grayscale)
  • Throughput requirements
  • Lossy vs. Loss less
  • Target technology
  • ASIC or FPGA
  • Memory

6
Trade Offs in a Design
  • Speed vs. Area (and Memory)
  • Standard allows for code-blocks to be
    entropy-coded/ decoded separately. So, parallel
    entropy coding engines can be employed to
    increase speed. This comes of course at the cost
    of extra memory and area
  • Speed vs. Optimal Bit Rate Control
  • The time spend in entropy coding can be reduced
    if less data are fed for entropy coding. This can
    be achieved by using higher quantization values.
    However, the bigger the quantization values the
    less optimization can be done by the bit-rate
    control algorithm

7
Trade-Offs in a Design, cont.
  • Programmability vs. Area
  • Hardwiring the DWT-filter type (5/3 or 9/7), the
    quantization tables, and the entropy coding
    switches can help to reduce the area requirements
  • Error resilience vs. Compression efficiency
  • Error resilience mechanisms introduce small
    overheads in the final stream size

8
The Need for a Reference Design
  • Allows for the developer to learn to manage the
    trade offs in a known design
  • Allows the developer to experiment with his/her
    own data
  • Allows the developer to identify the
    hardware/software structure of an optimal design

9
Why the Avnet Avalon Board
  • It offers a large, high performance FPGA with
    sufficient internal memory
  • It offers plenty of external memory expansion for
    non time critical buffering
  • It offers a fast host interface
  • Options for future expansion
Write a Comment
User Comments (0)
About PowerShow.com