Considering the stain variations, sometimes, it might be useful to extract the ECM information from the HE from the
(1) RGB/LAB space. Refer to another matlab implementation BDcreation_reg.m , it uses a more complex color segmentation (k-means in LAB) and decorrelation stretch, with manual RGB thresholds for nuclei. @Phil-Dua ,after porting the BDcreation_reg2.m, please consider the conversion of BDcreation_reg.m.
Considering the stain variations, sometimes, it might be useful to extract the ECM information from the HE from the
(1) RGB/LAB space. Refer to another matlab implementation BDcreation_reg.m , it uses a more complex color segmentation (k-means in LAB) and decorrelation stretch, with manual RGB thresholds for nuclei. @Phil-Dua ,after porting the BDcreation_reg2.m, please consider the conversion of BDcreation_reg.m.