Tamil OCR Conversion from Digital Writing Pad Recognition Accuracy Improves through Modified Deep Learning Architectures

Table 6

Performance evaluation of classifiers with accuracy for alphabet recognition.

References	Classifiers	Accuracy (%)
[24]	LSTM	90.3
[25]	Inception-v3	93.1
[26]	RNN	95.24
[27]	CNN	94
[28]	2C2S	93.25
Proposed	RTSBA with DWP-H dataset	98.7
	RTSBA with HP labs dataset	97.1

CNN, convolutional neural network; LSTM, long short-term memory; RTSBA, ResNet two-stage bottleneck architecture; RNN, recurrent neural network; 2C2S, two-channel and two-stream transformer; DWP-H, digital writing pad-handwritten.