Image analysis – Applications – Biomedical applications
Reexamination Certificate
1998-05-22
2001-12-11
Johns, Andrew W. (Department: 2621)
Image analysis
Applications
Biomedical applications
Reexamination Certificate
active
06330350
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
The present invention relates in general to a method and apparatus for automatically recognizing blood cells, and in particular, to a method and apparatus for automatically recognizing blood cell type by using an image processing and a neural network mechanism in a computer system.
2. Description of the Related Art
Normal white blood cells observed in blood are largely divided into neutrophils, lymphocytes, monocytes, basophils, and eosinophils. The occurrence of immature blood cells such as a blast, a normoblast, and an immature granulocyte is indicative of blood diseases. Accordingly, blood cell recognition methods are required to diagnose blood diseases such as leukemia. The recognition of blood cells can be performed manually or automatically. In the manual recognition method, an examiner has to observe blood cells with a microscope, so it requires a lot of time and effort. This method is therefore inappropriate to be utilized in large hospitals.
In counting the white blood cells in a blood sample, two methods have been developed and used for an automatic blood recognizer. The first method, which has been predominantly used to date, employs the combination of cytochemistry, electric impedance (impedance method), and a light scatter principle (optical method). The impedance method is based on the measurement of changes in an electrical current which are produced by a particle, suspended in a conductive liquid, as it passes through an aperture of known dimensions. In the optical method, a suspension of a diluted blood sample passes through a laser light test section where the cells scatter the laser light at different angles, yielding information about the size, internal structure, granularity and surface morphology. However, this method can only either recognize the five kinds of normal white blood cells or classify the cells into three or four separate functions. For instance, U.S. Pat. No. 4,581,223 introduces new staining solutions for an improved cytochemical reaction but only five individual white blood cells are identified by selective use of basic quaternary metachromatic dye staining of blood at a controlled temperature. Immature cells are not recognized. International patent application PCT/US88/00960 describes an electronic counting method but this method also enables recognition of the above five kinds of white blood cells. The patent application, PCT/US85/00840 is directed to a reference control solution for an electronic threshold setting, but only for three separate white blood cell control portions (lymphocytes, mononuclear cells and granulocytes). Other examples are also available. See European Pat. EP 0 424 871 A1 and U.S. Pat. No. 5,045,474. The most well known blood cell counters based on above principles are the so-called Coulter Counter and Cell-Dyn Counter.
The second method is based on a pattern recognition method in which the white blood cells are recognized after their images have been acquired followed by digitized and image-processed in a computer. U.S. Pat. No. 4,338,024 discloses a method of obtaining still images of objects in a flow stream on a CCD camera. Successive results of investigation of this patent were published in Clinical Chemistry (40/9, pp. 1850-1861, 1994). In U.S. Pat. No. 4,175,860, a method and apparatus is disclosed for use in performing automated classification of cells using a set of images, i.e., a high resolution image of primarily the nucleus of a cell and a low resolution image of the total cell. The high-resolution image is processed to obtain information on the texture whereas the low-resolution image is processed to obtain information on the size, density and color of the cytoplasm and the nucleus. In U.S. Pat. No. 5,123,055, a method and apparatus is provided for automatically identifying the five kinds of white blood cells. However, so far, to applicant's knowledge, there has not been a simple and efficient method of differentiating the five kinds of white blood cells and critical immature blood cells. In addition, none of above mentioned references have disclosed a method of recognition which combines both a statistical module and a neural network.
SUMMARY OF THE INVENTION
It is therefore one object of the present invention to provide a method and apparatus for accurately recognizing normal white blood cells as well as immature blood cells.
In accordance with one aspect of the present invention, a method and apparatus for automatically recognizing blood cells is provided. The method comprises obtaining image data of said blood cells, storing the image data of said blood cells in a memory, analyzing the image data of said blood cells and calculating a plurality of cell-characteristic parameter values for each of said blood cells, and recognizing each of said blood cells based on said plurality of cell-characteristic parameter values. In order to analyze the image data of blood cells, a group of nucleus pixels are first extracted from said stored image data and stored in the memory. The group of nucleus pixels are then segmented into individual nucleus clusters which represent each of the blood cells and the data of individual nucleus clusters are stored together with the associated cytoplasm data in the memory. Cell-characteristic parameter values for each of the blood cells are calculated based on its nucleus and associated cytoplasm data. The cell-characteristic parameter values are used in recognizing the blood cells.
REFERENCES:
patent: 4097845 (1978-06-01), Bacus
patent: 4175860 (1979-11-01), Bacus
patent: 4338024 (1982-07-01), Bolz et al.
patent: 4581223 (1986-04-01), Kass
patent: 5045474 (1991-09-01), Becker et al.
patent: 5123055 (1992-06-01), Kasdan
patent: 5541064 (1996-07-01), Bacus et al.
patent: 5741213 (1998-04-01), Kouchi et al.
patent: 5757954 (1998-05-01), Kuan et al.
patent: 5828776 (1998-10-01), Lee et al.
patent: 5911000 (1999-06-01), Shen
patent: 6151405 (2000-11-01), Douglass et al.
patent: 6169816 (2001-01-01), Ravkin
patent: 0 424 871 (1991-05-01), None
patent: WO 85/05450 (1985-12-01), None
patent: WO 88/07199 (1988-09-01), None
Harvey L. Kasdan, et al., Clinical Chemistry, vol. 40, No. 9, pp. 1850 to 1861, The White Iris Leukocyte Differential Analyzer for Rapid High-Precision Differentials Based on Images of Cytoprobe-Reacted Cells, Sep. 1994.
P. Joanne Cornblette, et al., Hematopathology, vol. 99, No. 1, pp. 72 to 81, “Evaluation of the Coulter STKS Five-Part Differential”, Jan. 1991.
Stuart A. Bentley, et al., Hematopathology, vol. 100, No. 6, pp. 626 to 632, “A Parallel Evaluation of Four Automated Hematology Analyzers”, Dec. 1993.
Malcolm L. Brigden, et al., Hematopathology, vol. 100, No. 6, pp. 618 to 625, “Evaluation of the Sysmex NE-8000”, Dec. 1993.
Wolfgang Huebl, et al., Hematopathology, vol. 103, No. 2, pp. 167 to 170, “Precision and Accuracy of Monocyte Counting”, Feb. 1995.
Ahn Hyo Sok
Grigoriev Andrei Y.
Kovalev Vassili A.
Myshkin Nikolai K.
Johns Andrew W.
Korea Institute of Science and Technology
Nakhjavan Shervin
Oblon & Spivak, McClelland, Maier & Neustadt P.C.
LandOfFree
Method and apparatus for automatically recognizing blood cells does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for automatically recognizing blood cells, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for automatically recognizing blood cells will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2599871