Figure 2:

Phase 2 of training, which represents the final model that was ultimately used to produce results from the validation set. The use of a recurrent neural network (RNN) with the gated recurrent unit (GRU) allows the three radiographs to be processed in series prior to the final output of whether the study is positive or negative for abnormality. The results from this model were reported and analyzed in the results section. CNN = convolutional neural network.