Fig. 4: Classification validation using fluorescent imagining. | npj Biosensing

Fig. 4: Classification validation using fluorescent imagining.

From: Realtime bacteria detection and analysis in sterile liquid products using deep learning holographic imaging

Fig. 4

a Hologram and corresponding fluorescent image samples showing our deep learning model detects and classifies E. coli (EC) and E. faecalis (EF) among generic PMs, as well as the detections below our confidence threshold (unknown). The figure also includes an in-focused closeup hologram for each type of particle showing their distinct diffraction patterns. b Receiver operator characteristic (ROC) curves and corresponding area under curve (AUC) values of our deep learning model for PMs, EC, and EF with vertical dashed lines marking 0.1 and 0.3% false positive rate (FPR), respectively. c The confusion matrix shows the accuracy and prediction errors of our classification for each type of particle evaluated at 0.1 and 0.35% FPR, respectively.

Back to article page