Lung segmentation on standard and mobile chest radiographs using oriented Gaussian derivatives filter

Table 5 Segmentation methods for comparison (for 247 images in JSRT database)

Reference	Method	Complexity	Overlap	Sensitivity	Specificity	Accuracy	Execution Time
Proposed method	Rule	Unsupervised and fully automated	0.870 ± 0.059	0.928	0.971	0.958	10-15 s (512 by 512)
[14]	Rule	Labelling	N/A	N/A	N/A	0.816	N/A
[34]	Model	Supervised	N/A	0.956	0.984	N/A	N/A
[17]	Hybrid: Model + pixel	Supervised	0.949 ± 0.020	N/A	N/A	N/A	N/A
[36]	Hybrid: Model + rule	Supervised	0.94 ± 0.053	N/A	N/A	N/A	N/A
[37]	Hybrid: Rule + Shape	Supervised	0.954 ± 0.015	N/A	N/A	N/A	85-90s (512 by 512)

ISSN: 1475-925X