Soft counting networks for bone marrow differentials

Abstract

Differential white cell counts from bone marrow preparations are very useful in evaluation of various hematologic disorders. It is tedious to locate, identify, and count these classes of cells, even by skilled hands. Automation of classification and counting would be of great benefit. However, the class structure of bone marrow or peripheral blood cells is not discrete; it represents a biological continuum of maturation levels. Because of this, there is uncertainty and overlap in characteristics of adjacent cell classes such that traditional pattern recognition techniques have difficulty in arriving at accurate cell counts. In this into one of the constituent groups and then simply increment the appropriate counter. However, 11s is true in many biological processes, the maturaticn of white blood cells represents a biological continiium; the transitions between categories are "artificial boundaries". Hence, many classification errors can, and do, occur between adjacent maturation cl,mses when crisp boundaries are assumed. In this paper, we use a soft computing approach to train neural networks to effectively produce the correct total class counts instead of focusing on classification and compare these results to several traditional classifiers. Myeloid Lineage ClassificationWe describe research efforts to perform standard classification of the 6 categories of granulocytic cells from bone marrow. Before automatic classification, single cells must be extracted, or segmented, from a microscopic bone marrow image and labeled. Then, we can extract features from each single ce!U and paper, we investigate soft counting networks that are trained to produce accurate overall class counts by allowing cells to have degrees of membership in multiple cell classes. This approach is applied to a bone marrow cell library and i s compared with other standard recognition algorithms.

Analysis of blood and bone marrow smears using multispectral imaging analysis techniques

“…In the past decades, there are only several groups working in it. Among them, both Pandit [14,15,16] and Keller's group [17,18,19,20,21] did a lot of meaningful work in the segmentation, feature extraction, and classification of the white blood cells in the grey or color bone marrow image. In the analysis procedures, the most important and difficult part is segmentation, but for a practical system, the reported segmentation results still need enhancement.…”

Precise Segmentation of Bone Marrow Cells by Using Multispectral Imaging Analysis Techniques

“…In the past decades, a lot of meaningful work had been done in bone marrow images segmentation [2][3][4][5][6][7][8][9][10][11]. Most of their work mainly focused on two difficult points: the segmentation of nuclei and the cytoplasm, along with cluster segmentation.…”