Optical Character Recognition Using Transformers |
Author(s): |
| G.Eshwar , SVS Group Of Institutions; L.Rishik Reddy, SVS Group Of Institutions; B.SaiVikas, SVS Group Of Institutions; Dr.B.Raghu, SVS Group Of Institutions |
Keywords: |
| Optical Character Recognition, Transformers, Natural Language Processing (NLP) |
Abstract |
|
Optical Character Recognition (OCR) is crucial for converting printed or handwritten text into machine-readable formats. It aids in document digitization, archival, and data extraction, enabling efficient conversion of visual textual information OCR automates the process reducing manual transcription time and error, enhancing information retrieval, and data analysis, and facilitating seamless integration of paper-based information into digital workflows. Transformers, a new architecture in Natural Language Processing (NLP), outperform older models like CNNs and LSTMs in tasks like text recognition from images (OCR) and machine translation. Their "self-attention" mechanism captures context better, and their parallel processing makes them faster. This project explores Transformers for Telugu-English & and English-Telugu machine translation and OCR, aiming for high accuracy and efficiency. |
Other Details |
|
Paper ID: IJSRDV12I20073 Published in: Volume : 12, Issue : 2 Publication Date: 01/05/2024 Page(s): 32-34 |
Article Preview |
|
|
|
|
