Mining images in biomedical publications: Detection and analysis of gel diagrams

Table 1 The results of the gel segment detection classifiers (top) and the gel panel detection algorithm (bottom)

	Method	Threshold	Precision	Recall	F-score	ROC area
		0.15	0.439	0.909	0.592
	Random forests	0.30	0.765	0.739	0.752	0.980
		0.60	0.926	0.301	0.455
Segments	Naive Bayes		0.172	0.739	0.279	0.883
	Bayesian network		0.394	0.531	0.452	0.914
	PART decision list		0.631	0.496	0.555	0.777
	Convolutional networks		0.142	0.949	0.248
Panels	Hand-coded rules		0.951	0.368	0.530

ISSN: 2041-1480