Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{156035,
        author = {Shruthi K.R and Abhishek and Bharath S and Pavan Kumar S and Madhu R},
        title = {Deep Learning Based Content Retrieval for Recognition and Classification in Historical Document},
        journal = {International Journal of Innovative Research in Technology},
        year = {},
        volume = {9},
        number = {2},
        pages = {571-577},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=156035},
        abstract = {This is quintessential because the variety of digitized historic files has expanded rapidly in latest decades. It affords environment friendly statistics retrieval and information extraction techniques to enable get entry to data. Such a technique to transform document images into written representations, it uses optical character recognition (OCR). At the moment, OCR methods frequently do not fit into the historical realm. In addition, they normally require a massive amount Annotated document. Therefore, this report will exhibit you some methods to enable OCR on historical data. Add some authentic, manually labelled coaching information to the photograph. Full featured OCR The device performs two main tasks: OCR and page layout analysis, which includes text block and line segmentation. Our segmentation method uses a recurrent neural network, while the OCR method is based on a fully convolutional network. Both strategies are cutting edge in the relevant field. built a new kind of Protonium Portal genuine dataset for OCR. All recommended techniques will be assessed in light of this data, which is freely available for research on this corpus. We display it with the aid of some real samples of annotated records, both segmentation and OCR jobs can be completed. The experiment goals to do this If your dataset is small, determine the satisfactory way to do it properly. We also show that the rating carried out is equal to or better than the scores of some contemporary systems. In conclusion, this study shows how to develop an effective OCR system for historical archives even in the absence of much training data.},
        keywords = {Deep Learning
Convolution neural network
Historical Document
Text Retrieval 
Optical Character Recognition
Deep Neural Network},
        month = {},
        }