Figure 1 | Scientific Reports

Figure 1

From: Nuquantus: Machine learning software for the characterization and quantification of cell nuclei in complex immunofluorescent tissue images

Figure 1

Nuquantus pipeline using example of a mouse cardiac tissue image.

(a) Original immunofluorescent image of mouse cardiac tissue obtained by confocal microscopy (scale 20μm). (b) Image preprocessing to correct and normalize illumination artifacts. (c) Extraction of image features including texture, shapes, local disorganization, size of the extracellular space and cell edges. (d) Nuclei selected by Nuquantus based on predictions made by a priori trained machine learning model. If some false positive nuclei were rendered in this step, they can be discarded by the user in the following optional step. (e) Optional user interaction step in which the user approves identified nuclei of cells of interest (cardiomyocytes)–circled in yellow. (f) Final output with classified nuclei of cells of interest (cardiomyocytes). (g) Quantitative output is provided with counted nuclei of different cell sub-groups.

Back to article page