Publication Date: 2023/07/17
Abstract: Text on an image often contains important information and directly carries high-level semantics in academic institutions and financial institutions. This makes it an important source of information and a popular research topic. Many studies have shown that CNN-based neural networks are very good at classifying images, which is the foundation of text recognition. By combining AI with the process of biometric identification, a technique for text recognition in academic institutions and financial institutions is performed using Convolutional Neural Network (CNN). Initially, preprocessing is done for making the document image suitable for feature extraction. One hot encoding- based feature extraction is performed. Two-dimensional CNN is used to classify the final features.
Keywords: Adam optimizer, AI, CNN, RMSprop, One hot encoding .
DOI: https://doi.org/10.5281/zenodo.8153358
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT23JUN2333.pdf
REFERENCES