Text detection and recognition through deep learning-based fusion neural network

<p>Text recognition task involves recognizing the text from natural image; it possesses various application, which aids information extraction through data mining street view like images. Scene two stages i.e., detection and recognition, in past several mechanisms has been proposed for accurate identification, these are either traditional approach or deep learning-based. All existing deep-learning methodology fails as this comprises character image data, further research develops an optimal architecture fusion neural network (FNN) identification recognition. FNN layers of convolutional well recurrent network. Within layer is utilized feature attaining classification prediction. Further, training established enhancement accuracy. Here Devanagari MLT-19 dataset evaluation FNN. Three different parameters considered during script word rate (CRR) (WRR). Further comparison with models performed to establish model efficiency shows observes 98.67% accuracy, 84.65% WRR 92.93% CRR.</p>