To be suitable for wireless multimedia applications, JPEG 2000 has to be robust to transmission errors. The core coding system, as specified in Part 1, defines error resilience tools to improve performances over noisy channels. However, these tools only detect where errors occur, conceal the erroneous data, and resynchronise the decoder. In other words, they do not correct transmission errors. Furthermore, these tools do not apply to the image header which is the most important part of the codestream. For these reasons, they are not sufficient in the context of wireless transmissions.

To address this issue, JPEG 2000 Wireless (JPWL) or Part 11 of the standard is standardising tools and methods to achieve the efficient transmission of JPEG 2000 imagery over an error-prone wireless network. More specifically, JPWL extends the elements in the core coding system described in Part 1 with mechanisms for error protection and correction. These extensions are backward compatible in the sense that decoders which implement Part 1 are able to skip the extensions defined in JPWL.

The JPWL system supports three functionalities:

• protection of the codestream against transmission errors,
  
• description of the degree of sensitivity of different parts of the codestream to transmission errors,
  
• description of the locations of residual errors in the codestream.

The JPWL system is illustrated in the figure below.


The error protection encoder and decoder consist in techniques to protect a codestream against transmission errors. The error protection encoding process modifies the codestream to make it more resilient to errors, e.g. by adding redundancy or by partitioning and interleaving the data. In particular, provisions are made to specifically protect the image header, as it is the most critical part of the codestream. Techniques to protect the codestream include Forward Error Correcting (FEC) codes, data partitioning and interleaving, and unequal error protection. The error protection decoding process detects the occurrence of errors and corrects them whenever possible.

The error sensitivity descriptor describes the degree of sensitivity of different parts of the codestream to transmission errors. This information is typically generated when the image is encoded using a JPEG 2000 Part 1 encoder, but it can also be directly derived from a Part 1 codestream. This information can subsequently be used when protecting the image. More specifically, sensitive parts of the codestream can be more heavily protected than less sensitive parts (unequal error protection).

The residual errors descriptor specifies the locations of residual errors in the codestream. The residual errors are the errors which cannot be corrected by the error protection decoder. This information is typically generated during the error protection decoding process. This information can subsequently be used in the JPEG 2000 Part 1 decoder to prevent decoding corrupted portions of the stream.

The work is also closely linked into the studies into JPEG 2000 security (JPSEC) in Part 8 of the standard, and to the client server communications aspects defined in Part 10 (JPIP).

Part 11 should become an International Standard (ISO/IEC 15444-11) in June 2007.

It is currently available for download from ITU-T as their Recommendation T.810

For further information, please contact:

Didier Nicholson, THALES Communications, chair of JPWL Ad hoc group.
or
Dr. Frederic Dufaux, EPFL, co-chair of JPWL Ad hoc group.