| Autor: |
Yokoyama, Kiyoko, Ishii, Naohiro, Suzumura, Nobuo, Naka, Ken-Ichi |
| Předmět: |
|
| Zdroj: |
Systems & Computers in Japan; Nov/Dec85, Vol. 16 Issue 6, p49-58, 10p |
| Abstrakt: |
One of the most important problems in biology is the morphological classification of cells as well as analysis of the relation between the morphology and function of the cell. Such analysis requires the extraction of feature parameters from cell images and classification based on their correlations. This paper seeks to design an image database whereby a number of cell images is stored in a computer for calculation of feature parameters and for the retrieval and automatic classification of images. A new algorithm for image processing is developed for acquisition of cell images. A procedure is developed to store cell images for calculation of feature parameters and automatic classification. In the past, attempts have been made to describe single cell images in accurate detail (e.g., by three-dimensional reconstruction of the cell image). However, those methods are not suited to the design of an image database in which a large number of cell images must be handled. From such a viewpoint, a method for morphological classification and feature analysis is developed whereby a cell image is represented by fewer feature parameters, making it possible to handle a large number of cell images. The method for acquisition of cell images, their storage and processing for feature extraction is systematically developed, arriving at a classification method based on those feature parameters. Factor analysis is applied to analyze the correlations among the feature parameters. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
Plný text