Title :
Semi-supervised color decomposition for histopathological images using exclusive component analysis
Author_Institution :
Bioinformatics Institute, Aegncy for Science, Technology and Research (A∗
Abstract :
In this paper, we propose a color decomposition method for hematoxylin-eosin (H&E) staining images using exclusive component analysis (XCA). We compare against two existing methods: conventional color deconvolution and color decomposition using non-negative matrix factorization method. By comparing the results, it shows that with the minimal effort from the user, the proposed method can not only produce good results, but also extract more information on the structure of cell organelles. Therefore, our proposed color decomposition method can provide more efficient color transformation for the further medical/biomedical image processing, such as segmentation, classification, etc.
Keywords :
"Image color analysis","Deconvolution","Integrated optics","Optical imaging","Optimization","Bioinformatics","Matrix decomposition"
Conference_Titel :
Machine Learning for Signal Processing (MLSP), 2015 IEEE 25th International Workshop on
DOI :
10.1109/MLSP.2015.7324359