Residual coding in document image compression

Abstract

Symbolic document image compression relies on the detection of similar patterns in a document image and construction of a prototype library. Compression is achieved by referencing multiple pattern instances ("components") through a single representative prototype. To provide a lossless compression, however, the residual difference between each component and its assigned prototype must be coded. Since the size of the residual can significantly affect the compression ratio, efficient coding is essential. In this paper, we describe a set of residual coding models for use with symbolic document image compression that exhibit desirable characteristics for compression and rate-distortion and facilitate compressed-domain processing. The first model orders the residual pixels by their distance to the prototype edge. Grouping pixels based on this distance value allows for a more compact coding and lower entropy. This distance model is then extended to a model that defines the structure of the residue and uses it as a basis for continuous and packet reconstruction which provides desired functionality for use in lossy compression and progressive transmission.

1-D chaincode pattern matching for compression of Bi-level printed farsi and arabic textual images

“…In the lossy compression case, all or some residual pixels are omitted and need not be encoded. In [17,16] a method of prioritization of the residual pixels for encoding has been proposed based on the distance of these pixels from their nearest edges. This prioritization helps to have an effective progressive compression since the most important information has already been encoded.…”

“…In the lossy or limited bit rate compression, all or some of the residual patterns may be omitted and hence need not be encoded. Since the residual patterns (and even the residual pixels) have not equal importance [17,15], it is reasonable to omit the less important ones and encode the rest. This technique preserves the reconstructed image quality as much as possible.…”

“…Of course, each pattern is not necessarily fully similar to its corresponding prototype, and is usually somewhat different. The difference image is called the ''residual pattern" or the ''error map", and its pixels are called the ''residual pixels" or the ''error pixels" [17]. The set of all prototypes is called the library.…”

Financial document image coding with regions of interest using JPEG2000

“…Many document image compression methods are proposed mainly for binary images, and partly for gray images: CCITT G4 (fax), JBIG1, JBIG2, and other methods [4,6,7]. As technology has improved, color becomes the norm for many document archives formats.…”

“…Consequently, several compression algorithms are proposed for coding color documents. Most good methods for document image coding, such as JBIG2, DjVu [1], and other methods introduced in [4,7,11], adopt hybrid-coding schemes, which decompose a document into several regions or layers, and each of these parts is coded by different compression methods.…”