Handwritten Bangla Character Recognition Using the State-of-the-Art Deep Convolutional Neural Networks

Table 3

The testing accuracy of VGG-16 Network, All-Conv Network, NiN, ResNet, FractalNet, and DenseNet on Digit-10, Alphabet-50, and SpecialChar-13 and comparison against other existing methods.

Types	Methods	Digit-10 (%)	Alphabet-50 (%)	SpecialChar-13 (%)
Existing approaches	MLP [7]	96.67	—	—
	MPCA + QTLR [56]	98.55	—	—
	GA [22]	97.00	—	—
	LeNet + DBN [57]	98.64	—	—
	VGG Net [9]	97.57	97.56	96.15
DCNN	All-Conv [10]	97.08	94.31	95.58
	NiN [11]	97.36	96.73	97.24
	ResNet [12]	98.51	97.33	97.64
	FractalNet [13]	98.92	97.87	97.98
	DenseNet [14]	99.13	98.31	98.18