Figure 2
From: Label-free imaging flow cytometry for analysis and sorting of enzymatically dissociated tissues
CNN for detection of cell aggregates. (A) Bright-field images of single cells and cell aggregates of human retinal organoid cells. Images are used to train a CNN for discrimination between single cells and cell aggregates. (B) Confusion matrix resulting when applying the CNN on the validation set. The validation accuracy is 80.3%. (C) Probability distribution resulting when applying the model to a testing dataset of dissociated Nrl-eGFP retina. Despite the different origin of the cells, the model is able to distinguish between single (left, low probability) and aggregated cells (right, high probability).
