Algum conteúdo deste aplicativo está indisponível no momento.
Se esta situação persistir, por favor entre em contato conoscoFale conosco & Contato
1. (WO2019025515) SYSTEM AND METHOD FOR COLOR DECONVOLUTION OF A SLIDE IMAGE TO ASSIST IN THE ANALYSIS OF TISSUE SPECIMEN
Nota: O texto foi obtido por processos automáticos de reconhecimento ótico de caracteres.
Para fins jurídicos, favor utilizar a versão PDF.

CLAIMS

What is claimed is:

1. A method for deconvoluting a multi-color channel image of a biological specimen that is stained by means of a staining system, wherein the image has undefined reference stain vectors and stain component images, the method comprising:

adopting prior knowledge of stain variability of the staining system as initial reference stain vectors;

based on the initial reference stain vectors, determining both the reference stain vectors and sta'm component images; and

based on the determined reference stain vectors and stain component images, deconvoluting the image into at least two-channel constituent sta'm component images.

2. The method of claim 1 , wherein adopting prior knowledge of stain variability includes assigning a Gaussian prior as a reference stain vector.

3. The method of claim 2, further including iteratively applying the Gaussian prior in two successive E step then M step, using an Expectation-Maximization (EM) model.

4. The method of claim 3, wherein the two successive E step and M step continue to be iteratively applied; and

upon attaining convergence, terminating the iterations.

5. The method of claim 4, further including allocating artificial latent variables, including defining an artificial latent variable residue (Resi) as a cross- product of a first stain vector and a second stain vector, in order to generate a vector that is independent of the first and second stain vectors.

6. The method of claim 5, wherein the first stain vector comprises a hematoxylin vector and the second stain vector comprises an eosin vector.

7. The method of any of claims 1 -6, wherein the multi-color channel image is deconvolved into a Hematoxylin (H-channel) image and an Eosin (E-channel) image.

8. A non-transitory computer-readable medium (105) including instructions which, when executed by one or more processors (102) of a tissue analysis system (100), cause the tissue analysis system (100) to perform the method of any one of claims 1-7.

9. A tissue analysis system (100) comprising one or more processors (102) and coupled to at least one memory (105) storing instructions which, when executed by the one or more processors (102), cause the tissue analysis system (100) to perform the method of any one of claims 1-7.