Table 5.

Comparative results on the Jacquard dataset.

Authors	Algorithm	Accuracy (%)
		IoU	SGT
Depierre et al. [6]	AlexNet	74.2	72.4
Morrison et al. [9]	GG-CNN2	84.0	85.0
Zhou et al. [14]	FCGN, ResNet-101	92.8	81.9
Zhang et al. [16]	ROI-GD, ResNet-101	93.6	-
Wang [55]	DD-Net, Hourglass	97.0	89.4
Yu et al. [57]	SE-ResUNet	95.7	-
	GR-ConvNet (b = 8, d = 0.0)	94.6	89.5
Ours	GR-ConvNet v2 (b = 16, d = 0.1)	95.1	91.4