Fig. 4.

Graphs illustrating the usage patterns of the datasets and models in selected studies. A: piechart with dataset distribution across the selected papers, B: piechart illustrating the model selection in the included articles. The total number of models is larger than the number of studies, as some studies trained more than one model. C: chronological chart of dataset usage focusing on inhouse/public dataset selection. The barchart illustrates the steady increase of studies using inhouse datasets.