Deep Learning in Historical Kannada Document
| Autor: | P. Ravi, Y. H. Sharath Kumar, C. Naveena |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
History
business.industry Deep learning General Chemistry Condensed Matter Physics computer.software_genre language.human_language Kannada Computational Mathematics language General Materials Science Artificial intelligence Electrical and Electronic Engineering business computer Natural language processing |
| Zdroj: | Journal of Computational and Theoretical Nanoscience. 17:4111-4115 |
| ISSN: | 1546-1955 |
| DOI: | 10.1166/jctn.2020.9028 |
| Popis: | Kannada is one of the famous ancient languages of the India and the official language of the State of Karnataka, which has a very large heritage. The digital analysis of these historical Kannada documents will provide us information about the culture and traditions that were practiced. Retrieving such information from paper documents, palm leaves or from stone carvings will enhance our knowledge. Investigating Historical document isn’t straight advance procedure because of low quality, differentiation, contrast and covering of characters. In this analysis, the authors propose a novel Scale invariant Feature Transform (SIFT) with deep learning classifier to recognize Historical kannada characters. To begin with, the character is divided utilizing Connected Component Analysis and later the Different SIFT Features are detached. At long last, form a powerful convolutional neural system classifier to recognize the Historical kannada archives. Proposed tale schemes during the preprocessing stage to guarantee strong, precise and constant grouping. They assess their strategy all alone datasets their characterization results surpass 97% on all datasets, which are superior to the cutting edge in this space. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|