Convolutional Neural Network Architectures to Solve a Problem of Tuberculosis Classification Using X-Ray Images of the Lungs

Table 4

Chest and spine detection results in 1000 images using MobileNet 1.0.

Images	Without detecting chest cage	Without detecting the spine
With pneumonia (500 images)	0%	24.8% (124 images)
Without pneumonia (500 images)	0%	11.2% (56 images)
Total error	0%	18% (180 images)