Intelligent Document Digitization and Classification using Hybrid Deep Learning and NLP Approaches
Fuente:
Zenodo
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Recurso digital |
| Publicado: |
Zenodo
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866901119954845696 |
|---|---|
| author | Fazliya S Evangline R Ms. Sumitha T |
| author_facet | Fazliya S Evangline R Ms. Sumitha T |
| contents | Handwritten archives are time-consuming, inclined to errors, and prone to bodily harm when stored, retrieved, and categorized manually. An AI-powered device for the digitization and categorization of handwritten police and criminal archives is proposed in this study. After enhancing scanned archives with an picture preprocessing pipeline, the device makes use of optical personality focus (OCR) to retrieve text. A Natural Language Processing (NLP)-based classification model is then used to manner the digitized text, routinely classifying files into preset businesses in a way scientific reports, witness statements, and FIRs. In order to facilitate equipped storage and high-quality retrieval, metadata extraction is used to pick out essential entities consisting of names, dates, places, and kinds of crimes. Unstructured handwriting inputs are transformed into safe, searchable, and analyzable digital statistics thru the integration of digitization and classification modules. Data-driven police work is made feasible with the aid of experimental assessment, which indicates extended accuracy in record categorization and textual content recognition, substantially slicing down on retrieval time. This work gives a scalable, multilingual, and impervious answer for digital policing projects, assisting to modernize the administration of regulation enforcement records. |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19221262 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Intelligent Document Digitization and Classification using Hybrid Deep Learning and NLP Approaches Fazliya S Evangline R Ms. Sumitha T Document Digitization Handwriting Recognition OCR NLP Document Classification. Handwritten archives are time-consuming, inclined to errors, and prone to bodily harm when stored, retrieved, and categorized manually. An AI-powered device for the digitization and categorization of handwritten police and criminal archives is proposed in this study. After enhancing scanned archives with an picture preprocessing pipeline, the device makes use of optical personality focus (OCR) to retrieve text. A Natural Language Processing (NLP)-based classification model is then used to manner the digitized text, routinely classifying files into preset businesses in a way scientific reports, witness statements, and FIRs. In order to facilitate equipped storage and high-quality retrieval, metadata extraction is used to pick out essential entities consisting of names, dates, places, and kinds of crimes. Unstructured handwriting inputs are transformed into safe, searchable, and analyzable digital statistics thru the integration of digitization and classification modules. Data-driven police work is made feasible with the aid of experimental assessment, which indicates extended accuracy in record categorization and textual content recognition, substantially slicing down on retrieval time. This work gives a scalable, multilingual, and impervious answer for digital policing projects, assisting to modernize the administration of regulation enforcement records. |
| title | Intelligent Document Digitization and Classification using Hybrid Deep Learning and NLP Approaches |
| topic | Document Digitization Handwriting Recognition OCR NLP Document Classification. |
| url | https://doi.org/10.5281/zenodo.19221262 |