Researchers have developed new methods for optical character recognition (OCR) specifically for handwritten Bangla texts. One study focuses on word recognition using deep learning models like SSD with MobileNetV2 and Faster R-CNN with InceptionResNetV2, achieving an F1-score of 92.61% with an ensemble model. The second study addresses color-independent word segmentation from transcribed Bangla passages, reporting a 91.20% F1-score and highlighting its applicability to smartphone-captured images. AI
IMPACT Advances OCR capabilities for under-resourced languages, potentially improving accessibility and data extraction.
RANK_REASON The cluster contains two arXiv papers detailing new research in OCR for Bangla texts.
- alphaXiv
- arXiv
- Bangla
- CatalyzeX
- DagsHub
- Faster R-CNN
- Gotit.pub
- Hugging Face
- InceptionResNetV2
- MobileNetV2
- Non-Maximum Suppression
- optical character recognition
- ScienceCast
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →