Use of error image for unmixing artifact removal in linear spectral unmixing

摘要

A method of removing a potential false positive result in a stained sample. The method includes the steps of: acquiring a spectral image set of a stained sample; performing linear spectral unmixing of the spectral image set to obtain a plurality of rule images; calculating an error image for the stained sample; thresholding at least one of the plurality of rule images, wherein thresholding is based on the error image; and removing a potential false positive result from the at least one of the plurality of rule images based on thresholding.

主权项

1. A method of removing a potential false positive result in a stained sample, the method comprising the steps of:
acquiring a spectral image set of a stained sample; performing linear spectral unmixing of the spectral image set to obtain a plurality of rule images; calculating an error image for the stained sample; thresholding at least one of the plurality of rule images, wherein thresholding is based on the error image and further comprises segmenting the error image to create a high error mask image and a low error mask image; and separately applying both mask images to at least one of each of the plurality of rule images; and removing a potential false positive result from the at least one of the plurality of rule images based on thresholding.