|
Work item:
|
H.BWC
|
|
Subject/title:
|
Biomedical and general waveform signal coding
|
|
Status:
|
[Carried to next study period]
|
|
Approval process:
|
AAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
-
|
|
Liaison:
|
DICOM WG32
|
|
Supporting members:
|
Fraunhofer HHI, Dolby Laboratories
|
|
Summary:
|
Currently, there lacks a codec with sufficient compression capability for neurophysiology and electrocardiography data and other similar types of biomedical waveform signals. Current audio codecs have significant limitations when applied to biomedical waveform data due to the application of psychoacoustic masking and limitations on channel number and block size. As such, a new standard that can support efficient lossless and near-lossless compression and transmission specifically for biomedical waveform data is almost certainly needed.
Neurophysiology data consists of 16-24 bit data with between 21 and several hundred channels sampled at between 256 and 40K Hz. Clinical scalp EEG is typically 21 channels sampled at between 256 and 1000 Hz. Clinical intracranial EEG consists of between 64 and several hundred channels sampled with a bit depth of 16 or 24 bits at around 2000 Hz. Research on intracranial EEG (human and animal neurophysiology) involves up to several hundred channels sampled at a bit depth of 16 or 24 bits and a sampling rate up to approximately 40K Hz.
The codec will need to meet the following minimum requirements:
1. Be able to support 16 to 24 bits (possibly higher) per sample.
2. Be able to support a large range of sampling rates (500–2000 sample/s for clinical use, up to 40k samples/s for research).
3. Be able to support a large number of channels (up to hundreds).
4. Be able to support channels with the same sampling rates, with support for channels with different sampling rates being potentially considered as well.
5. Be able to support lossy, near-lossless and lossless compression. A lossless-only approach will not produce better than a 2-3X compression ratio, which is not sufficient.
6. Be able to support blocking and indexing for rapid access within large datasets. Block size should be optimized, and an index to blocks can be stored outside of encapsulated bitstream (such as in fragments/frames encoding in DICOM).
7. Have a mode where independent decoding of channels is supported.
8. Associated metadata should be supported (possibly stored outside of the bitstream).
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2024-06-21 14:16:31
|
|
Last update:
2024-06-21 14:31:31
|
