Automatic Pattern Analysis Research Articles

Can seismic stratigraphy problems be solved using automated pattern analysis and recognition?

The Problem: Four wells had been drilled in the Four Horses area of the Powder River Basin in Weston County, Wyoming. Interpretation of well logs indicated that the sand content in the formation varied widely. The key question in the development of the area was the locations of the boundaries between the different sand contents.

Automated pattern analysis of growing rat lung.

This study discusses the use of automatic image analysis equipment in quantifying the structural changes during growth of rat pulmonary parenchyma. At birth the latter consists of primary septa delineating a coarse structure of alveolar sacs which are then subdivided into alveoli by outgrowth of secondary septa from day 4 onwards. This process is followed by internal restructuring of the septa in the third week which further increases the surface area by formation of small irregularities. A partial quantification of these processes was obtained by measuring the full set of parameters available with the Quantimet 720 using light microscopy on rat lung sections at the ages of 4, 10 and 21 days. This yielded basic data, e.g. volume and surface densities of air space and tissue. Using closure and opening as preprocessing operations made it possible to split up such parameters into contributions from different ranges of detail resolution. The mass of eighty-four variables obtained could be reduced to six independent factors by Factor Analysis techniques. Interpretation through their correlation with the original variables yielded septal surface, septal volume, minor and major axes of the large air spaces, and surface detail as important structural descriptors explaining 70% of total variance. The remainder is attributed to between-field variations.

Preface to the special issue

THERE are several trends that make the use of automated pattern analysis in medicine and biology of potentially great importance.

Automated Pattern Analysis of Corneal Endothelium

A proper understanding of the endothelial cell morphology of the cornea is of great significance to the corneal surgeon. Specular microscopy allows direct visualization of endothelial cell morphology and a proper analysis of these data needs automated computerized type of systems. Automated pattern analysis seems to offer a good option in this direction. From our experience it appears that such systems can be used with great advantage in the analysis of endothelial cell morphology. The data obtained from such analysis help to develop some indices for various facets of endothelial cell morphology.

Quantitative structure‐activity correlation of pyrethroids against Musca domestica and Blattella germanica

AbstractStructure‐activity correlations of pyrethroids to Musca domestica and Blattella germanica were studied in terms of partition coefficient, electronic effect and conformation. Relative Rf values of 12 pyrethroids measured by thin‐layer chromatography were related to the relative partition coefficients and electronic effect. The relative size and shape of six pyrethroids were determined by probabilistic automatic pattern analysis using Corey‐Paulings' molecular models. There is no correlation of knockdown of M. domestica and partitioning. The simple correlation, with 0.10< P <0.01 of partitioning, electronic effect and size and shape of pyrethroids to toxicity were 0, 0.2–0.3, 0.7–0.8 respectively while thermodynamic treatment gave corresponding correlations of 0, 0.4, 0.6.

Recognition of printed chinese characters by automatic pattern analysis

An approach to pattern recognition by computer, using analysis of pattern structure, isexplored. The approach is programmed and tested on a set of Chinese characters. The input to the program is a black-white matrix of points depicting a single character. The program produces a description of the character on two levels: (i) the internal structure of each connected part of the character, and (ii) the arrangement in two dimensions of the connected parts. The description is achieved by producing a structural representation of the character. The structure corresponding to level (i) is a graph, the edges of which correspond to parts of strokes. The structure corresponding to level (ii) is a tree, whose terminal elements correspond to connected parts of the character and whose nodes correspond to geometric relations. A method has been devised for producing a numeric code for each character. The codeis generated from the structural representation of a character, and is used for recognition.