Skip to main content

Bit-droppable Facsimile Encoding for Packet-Switched Networks.

01 January 1989

New Image

We present an image compression algorithm that is suitable for transmitting a high-resolution (400 dot-per-inch) facsimile image over a packet network. Its main features are: (a) It decomposes the image into non-droppable (essential) and droppable (supplementary) bits. (b) The non-droppable bits describe the image at low resolution (200 dpi) and are coded with the standard CCITT group III/IV algorithm. (c) The droppable (supplementary) bits are approximately 45% of the total number of bits and can be dropped by the network to relieve congestion. (d) It is information preserving, assuming that all droppable bits are available. (e) Its compression efficiency is about equal to that of direct G3/G4 CCITT encoding of the high resolution (400 dpi) image. (f) It does not need frame memory. This is the first algorithm for bi-level images that separates the image into droppable and non-droppable information while maintaining an acceptable compression efficiency. The separation of information into droppable and non-droppable is crucial for efficient congestion control in a packet network carrying simultaneously non-bursty, delay-sensitive (voice, video) traffic and bursty, delay-tolerant (computer data, facsimile) traffic. An additional advantage of this data representation is its compatibility with the large base of currently installed facsimile terminals. If our scheme for decomposing an image into high-resolution and low-resolution components is applied recursively, what results is an encoding algorithm that allows the progressive refinement of the encoded image into higher and higher resolution. Such progressive presentation would be desirable for CRT viewing of an image being delivered over a medium-rate channel.