Figure 3.

Deep learning provides accurate and rapid approach for screening drug and nanocarrier drug system. a) Confusion matrices for deep learning for classification of four classes. IE, LE, ME, and HE present ineffective, low, medium, and high efficacy, respectively. Color bar presents the accuracy. b) Accuracy of each test cell lines in a different set of models. X‐axis presents chosen training data of each model and color bars present the test data. K, R, G, and C present bright‐field, Annexin V‐APC, anti‐EGFR‐FITC, and self‐fluorescence of Cur channel, separately. c) Accuracy of LDH‐VP16, VP16, and untreated group in A549‐KRG model at 2 h/6 h under different stages of efficacy. IE, LE, ME, and HE present ineffective, low, medium, and high efficacy, respectively.