Table 1.

Segmentation accuracy and obstacle distance estimation error for obstacle datasets.

		Proposed Method	Conventional Method (Modified Version of [23])
Segmentation accuracy	Precision	81.4%	57.5%
	False positive rate	5.9%	14.2%
	Recall	74.4%	37.6%
Obstacle distance estimation error	Average error	1.6 cm	9.9 cm
	Standard deviation of error	5.8 cm	11.4 cm