Handwritten Character Recognition from Image Using CNN

Handwritten character identification has always been an intriguing area of study in the realm of pattern recognition in image processing. Because of its numerous applications, the requirement for identifying handwritten characters is growing every day. Many scholars have defined their work in this field, and additional research is being carried out to obtain high precision. In compared to other major languages such as Bangla, there are numerous works in handwritten character recognition available for English. The goal is to present a comprehensive, effective, and efficient method for classifying and recognizing both Bangla and English letters. An extended convolution neural network (CNN) model has been suggested to recognize Bangla and English characters. Character recognition is achieved through segmentation, feature extraction, and classification. CNNs were recently discovered to be adept at English text detection. A CNN-based Bangla handwritten character recognition system is also being researched. A total of 23,040-character samples have been used, with 25% of the data having served as a test set and the remaining 75% having been used to train the recognition model.