Elsevier Science Home
Computer Physics Communications Program Library
Full text online from Science Direct
Programs in Physics & Physical Chemistry
CPC Home

[Licence| Download | New Version Template] adcr_v1_0.gz(54 Kbytes)
Manuscript Title: Packing and depacking histograms with statistical processing.
Authors: S. Louvel, J.-F. Chamayou
Program title: compression and lissage
Catalogue identifier: ADCR_v1_0
Distribution format: gz
Journal reference: Comput. Phys. Commun. 93(1996)303
Programming language: Fortran.
Computer: TERMINAL-X.
Operating system: UNIX.
RAM: 3K words
Word size: 32
Keywords: General purpose, Statistical methods, Histogram, Image processing, Empirical function, Spline.
Classification: 4.13.

Nature of problem:
The storage of a great number of histograms can be an expensive operation in terms of memory size. This program allows the packing and the depacking of histograms without a too large loss of information. The rate of compresion is about 1 to 10 (identifier label included) and the method used holds some statistical properties satisfying some non parametrical tests.

Solution method:
The principle of packing is based on a statistical processing of data using the "cumulated histograms method". Thus, the information carried by the histogram is restricted in the jumps abscissae of the empirical function rather than in the observed bin size of the histogram. The result of this compression will be a set of S pairs of ASCII characters who produces successively the jumps abscissae and its number of repetitions. After the decompression, to make up for the lost information, we use a spline smoothing of the restored empirical function to estimate, with high precision, the original histogram.

After compression, the histogram looks like S pairs of integers (number of repetitions, jumps abscissae) coded into (2 x S) ASCII characters. The best results are obtained when the quantity (2 x S) is included between 7 and 10% of the number of histogram bins. This comes from an empirical research from a lot of tests.

Unusual features:
The aim of this program is the compression of gray levels histograms, with a natural extension in the multidimensional histograms (RED, GREEN, BLUE). Other subprograms used: S-plus software (optional), version 3.1, Statistical Sciences Inc. and NAG library.

Running time:
The computational work to account is divided into three parts: the compression, the decompression and the smoothing. Each one of these operations take negligible time, but the ture goal of this program is to reduce data storage requirements.