Customer Reviews

Recursive Block Coding for Image Data Compression

5 star:		(0)
4 star:		(1)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

Data compression methods compress binary bits or objects made up of binary bits.

But this book discusses an approach to compressing energy into a new space, a space where data consists of varying degrees of somewhat disposable abstract clusters of energy.

Therefore, an approach to a model for recursive energy compression prior to image data compression is actually the topic discussed in this book not any form of lossless recursive data compression, especially computable reducibility models. As I will explain, the books focus is not really to discuss a specific data compression method.

Watch your HDTV picture long enough and eventually you will see tiny blocks appear and mess up your crystal clear picture.

History records that in the 1980's, this author, Paul Michael Farrelle, tried a new approach to conventional cosine transform coding aimed at minimizing those little block boundary problems that cause those tiny blocks.

It is public record that in the 1980's, Farrelle et al, discussed a new subordinate approach, as an addition to the prior art, utilizing conventional cosine transform coding of image data, non adaptive zonal transform coefficient coding and including uniform quantizers. As may also be evident by your television experience, it is common knowledge that even the many revisions of this configuration proved to be limited.

His new approach was designed to minimize errors which occur at the edges of those tiny blocks that mess up your crystal clear picture when using conventional cosine transform coding. Farrelle's error correction approach compares compression results of the block edges using the exact same uncompressed data.

This book includes Adaptive Coding as well as describes and elaborates on the type of Transform Coding, in which information is discarded, thereby compressing 50 to 1 (Page 3, chapter 2). In this context reduction is not synonymous with reducibility. The reason is that the term reducibility implies a degree of intact incompressibility (Chaitin Irreducibility or more abstractly Chaitin Kolmogorov Complexity) and if that independent existence could be achieved by discarding information, then it would require the use of the (in my opinion) discredited counting argument models espoused by David Salomon in his (error prone) data compression reference/handbook series. I am saying that I think that the counting argument models are irrational. One fine aftereffect of this information altering paradigm is the exercise for coding into Braille which should be accomplished without losing information, but that is not apparent in Salomon`s models. (Handbook of Data Compression, 5th edition, Exercise for Section 1.1.1 - Braille, on page 26, exercise box #1, when referring to information take note of the word "most" in the exercise.) So the connotations as well as the implications of calling these models computable reducibility models which are analogous to recursive data compression are far reaching and inaccurate.

(c)