Corrigendum

This article refers to:

Road network extraction: a neural-dynamic framework based on deep learning and a finite state machine

Wang, J., Song J., Chen, M., and Yang, Z. 2015. “Road Network Extraction: a Neural-Dynamic Framework Based on Deep Learning and a Finite State Mmachine.” International Journal of Remote Sensing 36 (12): 3144–3169. http:/dx.doi.org/10.1080/01431161.2015.1054049

The authors would like to inform the readers that the quality index in the original paper are calculated using a false equation. These errors have been corrected in following tables. The results do not affect the main conclusions in the original paper.

We apologize for any inconvenience this error may have caused.

Table 1. Quantitative evaluation for the four test datasets shown in Figure 13.

Dataset	Resolution (m)	Completeness (%)	Correctness (%)	Quality (%)
Dataset-1	0.1	74.4	83.2	64.7
Dataset-2	0.2	80.8	81.4	68.2
Dataset-3	1.0	87.2	93.3	82.1
Dataset-4	1.0	91.4	93.1	85.7

Table 2. Performance of the road network extraction from the four test cases shown in Figure 15 and 16.

Image	Completeness (%)		Correctness (%)		Quality (%)
	Object-based	Proposed	Object-based	Proposed	Object-based	Proposed
Test case 1	60	78	71	76	48	63
Test case 2	95	93	80	96	76	90
Test case 3	93	92	85	95	80	88
Test case 4	70	96	84	96	62	92

Table 3. Comparison and quantitative evaluation for complex urban areas shown in Figure 17.

Method	Completeness (%)	Correctness (%)	Quality (%)	No. of parameters	Processing time (s)	Interaction required
Object-based	60	64	45	13	> 300	Parameter turning
Poullis (2014)	68	64	62	2	≈ 86	None
Proposed	75	70	57	None	< 30	Seed point