Classification of immature and mature cells of the neutrophil series using morphometrical parameters.

Quantitative morphological data of six classes of immature and mature cells of the neutrophil series of the bone marrow of normal persons were used for statistical classification experiments (myeloblasts, promyelocytes, myelocytes, metamyelocytes, bands and segments). On each cell, parameters were measured directly from the image or calculated from the shape of the density histogram or the counting densitogram using a Texture Analysis System (E. Leitz, Wetzlar, Germany). The parameters were analyzed with the interactive statistical pattern recognition system ISPAHAN. One half of the data were used as a learning set and the other half as the test set. The parameters were compared according to their performance in discrimination between the classes, alone and in combinations. Parameters not contributing to an improvement of the discrimination were disregarded. Eleven parameters were selected and used for classification by two different methods: a stepwise and a "one-shot" method. Stepwise classification resulted in a 79% correct classification rate. Most errors occurred between cell classes in neighboring stages of maturation. In 96% of all cases the computer classification was either in accordance with that of the technician or with a cell class of a neighboring maturation stage. One step classification by the computer was in agreement with the technicians in 82% of the cases. For 98% of the cells the computer classification was either in accordance with that of the technician or with a cell class of a neighboring maturation stage. The data set was collected by two technicians, operating independently. Differences in their interpretation of the maturation stage were found by comparing the performance of classifiers based on both cell samples. Since the images of the cells were not available for reexamination, the causes of disagreement in classification between the technicians and between computer and technicians could not be evaluated.

Relevance of morphometric parameters for the classification of normal white blood cells.

An interactive morphometry system (Leitz T.A.S.) was used to extract 39 parameters from 1506 cells of five normal WBC classes. The data set was split in a learning set and a test set, and the value of each single parameter and of combinations of parameters was assessed with the interactive pattern recognition system ISPAHAN. A combination of nine densitogram parameters gave a correct classification in 92.6%; adding four geometrical and seven counting densitogram parameters improved this to 98.5%. It appeared that simple measurements are sufficient for a correct differentiation into the normal white blood cell types. In a pilot study fo 40 normal myeloid bone marrow cells of four classes a correct discrimination between mature and immature cells was reached in 93%. Extraction of parameters with the T.A.S. and their subsequent evaluation with ISPAHAN appeared to be an efficient method of white blood cell classification probably including immature cell types.

The use of ISPAHAN: interactive system for statistical pattern recognition and analysis.

ISPAHAN, the interactive system for statistical pattern recognition and analysis, was developed at the Department of Medical Information at the Free University of Amsterdam. It has been used in many pattern recognition problems, such as white blood cell recognition, typification of wave forms in ECG analysis, segmentation of ECG signals and resonance detection in high-energy particle physics. The structure and capabilities of ISPAHAN are presented along with an example of its use in the field of white blood cell recognition.

Texture of white blood cells expressed by the counting densitogram.

The counting of particles and holes in white blood cells during thresholding at different gray levels results in a histogram, the counting densitogram, which contains information about granularity. Parameters describing the shape of this histogram appear to be sufficiently characteristic for the five normal white blood cell classes to allow a 84% correct discrimination between them. This method to quantitate granularity could be valuable for the analysis of cell texture and therefore, when used in combination with geometrical and density histogram parameters, could contribute to the results of morphometric discrimination between other than normal white blood cells.

Demonstration of segmentation techniques for EEG records.

In this paper three different techniques for segmenting EEG's are presented. The principles of these techniques (the Kalman filter approach, the power spectrum analysis and the texture matrix approach) are explained and the results obtained summarised. The segmentation is used in an interactive EEG interpretation system.