Open Vocabulary Word Recognition From Transcribed Bangla Texts
作者: Faias Satter, Sk. Md. Masudul Ahsan
分类: cs.CV, cs.LG
发布日期: 2026-10-01 (更新: 2026-10-02)
备注: 6 pages, 4 figures, 5 tables. Accepted version of the paper published in the 2023 26th International Conference on Computer and Information Technology (ICCIT). Code: https://github.com/FaiasPromit/Open-Vocabulary-Word-Recognition-From-Transcribed-Bangla-Texts.git
期刊: 2023 26th International Conference on Computer and Information Technology (ICCIT), Cox's Bazar, Bangladesh, 2023
DOI: 10.1109/ICCIT60459.2023.10441393
💡 一句话要点
提出深度学习模型以解决孟加拉手写文本识别问题
🎯 匹配领域: 支柱三:空间感知与语义 (Perception & Semantics)
关键词: 光学字符识别 深度学习 目标检测 手写文本识别 孟加拉文 集成模型 非极大值抑制
📋 核心要点
- 现有的OCR系统在处理孟加拉文时缺乏可靠的解决方案,尤其是手写文本的识别效果较差。
- 本研究提出了基于深度学习的手写孟加拉单词识别方法,使用了SSD、Faster R-CNN及其集成模型。
- 实验结果表明,集成模型在F1-score上达到了92.61%,并且单词识别率为96.12%,显示出显著的性能提升。
📝 摘要(中文)
光学字符识别(OCR)技术能够扫描纸质文档并提取文本,极大地方便了人们的工作。然而,现有的OCR系统在处理孟加拉文时仍面临诸多挑战,尤其是手写文本的识别。本研究旨在通过深度学习技术识别手写孟加拉单词图像。研究中使用了三种目标检测模型:基于MobileNetV2的SSD、基于InceptionResNetV2的Faster R-CNN,以及这两者的集成模型。为提高模型效果,提出了一种改进的非极大值抑制方法。通过对9841幅多样化手写风格的孟加拉单词图像进行训练和测试,集成模型在F1-score上达到了92.61%的优秀表现,单词识别率为96.12%。
🔬 方法详解
问题定义:本研究旨在解决手写孟加拉文本的识别问题。现有的OCR技术在处理此类文本时效果不佳,尤其是在多样化的手写风格下,识别准确率低。
核心思路:论文提出利用深度学习模型进行手写单词图像的识别,采用SSD和Faster R-CNN两种目标检测模型,并结合集成学习的方法,以提高识别精度。
技术框架:整体架构包括数据集的构建、模型的训练与测试,以及后处理阶段。研究中使用了9841幅手写孟加拉单词图像作为训练数据,模型经过训练后进行性能评估。
关键创新:最重要的创新在于引入了改进的非极大值抑制方法,以提升模型的识别效果。这一方法有效减少了误识别的情况,提高了整体性能。
关键设计:在模型设计中,SSD与Faster R-CNN的网络结构经过优化,损失函数也进行了调整,以适应手写文本的特征。此外,集成模型的设计使得两种模型的优点得以结合,进一步提升了识别准确率。
🖼️ 关键图片
📊 实验亮点
实验结果显示,集成模型在F1-score上达到了92.61%,在单词级别的识别率为96.12%。相比于单独模型,集成模型在性能上有显著提升,展示了深度学习在手写文本识别中的有效性。
🎯 应用场景
该研究的潜在应用领域包括教育、文档数字化和语言处理等。通过提高孟加拉手写文本的识别准确率,可以为相关行业提供更高效的文档处理解决方案,促进信息的数字化和可访问性,具有重要的实际价值和社会影响。
📄 摘要(原文)
An optical character recognition (OCR) can scan a paper and extract text using technology, making people's jobs easier. While various OCR systems are available in the software industry, finding a reliable equivalent solution for Bangla takes much work. When it comes to handwritten texts, the situation is much more unusual. Recognizing words from word images is the most critical stage in any OCR process. It is the second stage after segmenting words from text pictures. If this stage fails, the overall performance of the OCR will be poor, regardless of how well the other phases perform. This study aims to recognize words using deep learning in a handwritten Bangla word image. Three object detection models, SSD with MobileNetV2, Faster R-CNN with InceptionResNetV2, and an ensemble model of these two, have been used to train and test handwritten word images. A modified Non-Maximum Suppression has been introduced to enhance the effectiveness of the models' results. A customized dataset of 9841 handwritten Bangla word images has been compiled, featuring diverse handwriting styles from various individuals. All three models' performances have been checked against the test dataset, and the ensemble model has been the most impressive, with an F1-score of 92.61%. Also, at the word level, the ensemble model correctly recognizes 96.12% of the words to some extent. The system can be further improved by introducing a post-processing phase to correct errors generated by the system.