BengaliNet: A Low-Cost Novel Convolutional Neural Network for Bengali Handwritten Characters Recognition

As it is the seventh most-spoken language and fifth native in world, domain of Bengali handwritten character recognition has fascinated researchers for decades. Although other popular languages i.e., English, Chinese, Hindi, Spanish, etc. have received many contributions area recognition, not noteworthy this because complex curvatures similar writing fashions characters. Previously, studies were conducted by using different approaches based on traditional learning, deep learning. In research, we proposed a low-cost novel convolutional neural network architecture characters with only 2.24 to 2.43 million parameters number output classes. We considered 8 formations CMATERdb datasets previous training phase. With experimental analysis, showed that our system outperformed works margin all datasets. Moreover, tested trained models available such as Ekush, BanglaLekha, NumtaDB Our achieved 96–99% overall accuracies these well. believe will be beneficial developing an automated high-performance tool