Fig. 3: The number of images used for model testing split by image class. | Nature Machine Intelligence

Fig. 3: The number of images used for model testing split by image class.

From: Common pitfalls and recommendations for using machine learning to detect and prognosticate for COVID-19 using chest radiographs and CT scans

Fig. 3

It is noted that we exclude Bai et al.81 and Zhang et al.27 from the figure as they used far more testing data (14,182 and 5,869 images respectively) than other papers. There were a large number of images (1,237) in the testing dataset in Wang et al.54 that were unidentified in the paper (we include these in the unspecified COVID-19 negative).

Back to article page