Skip to main content
Fig. 6 | Biological Procedures Online

Fig. 6

From: Automated macrophage counting in DLBCL tissue samples: a ROF filter based approach

Fig. 6

Segmentation methods (S1) - (S4) coping with problems in tissue preservation. Features have been identified based on information from all of three analyzed channels. adSample with vessel (bottom left) and erythrocytes (middle).a — Original single-channel image (whole tissue sample No. 24, cutout from tile No. (42, 16), CD163 +/555 nm), contrast enhanced by factor 1.5, scale bar 45 μm. b — Result of (S1), (S2); vessel as a hyperfluorescent feature removed, erythrocytes partly ignored. c — Result of (S3), contrast enhanced by factor 1.5; vessel erroneously marked as target area, erythrocytes partly ignored. d — Result of (S4), contrast enhanced by factor 1.5; vessel as well as erythrocytes erroneously marked as target area. ehSample with tissue fold.e — Original single-channel image (whole tissue sample No. 31, cutout from tile No. (3, 20), CD14 +/488 nm), contrast enhanced by factor 2, scale bar 45 μm. f — Result of (S1), (S2); fold as a strongly fluorescent feature removed, macrophages under the fold partly detected. g — Result of (S3), contrast enhanced by factor 2; fold erroneously marked as target area. h — Result of (S4), contrast enhanced by factor 2; fold erroneously marked as target area. ilSample with staining artifact (splatter of staining liquid).i — Original single-channel image (whole tissue sample No. 11, cutout from tile No. (18, 21), CD163 +/555 nm), scale bar 45 μm. j — Result of (S1), (S2); splatter as a hyperfluorescent feature removed, macrophages close to its border properly detected. k — Result of (S3); splatter erroneously marked as target area. l — Result of (S4); splatter erroneously marked as target area

Back to article page