Lung Diseases Detection Using Various Deep Learning Algorithms

Table 2

Accuracy of our model with existing works related to tuberculosis.

Authors	Dataset	Techniques	F1-score	Accuracy	Recall	Precision
Momeny et al. [25]	Annotated (1078 images)	CNN	89	93	85	92
Lopes and Valiati [26]	ImageNet (10000 images)	CNN	—	76	—	—
Sineglazov et al. [27]	National institute of phthisiology and pulmonology (9311 slices)	CNN	99.04	99	97.6	98.34
Mamalakis et al. [28]	CXR dataset (3883 images)	ResNet CNN	81.64	71	91	78.6
Duong et al. [29]	(i) Montgomery county & CXR dataset (138 images); (ii) Shenzhen dataset (662 images)	CNN	97.9	99	—	—
Our proposed model	Chest X-ray (662 images)	CNN	97.99	99.4%	98.8	98.55