Jul 2026· IEEE Jordan Conference on Applied Electrical Engineering and Computing Technologies· pp. 544-549· 0 citations· 23 references
Abstract
Scene text detection and recognition in multilingual environments remains challenging, particularly for morphologically complex scripts such as Arabic. This paper presents an end-to-end deep learning system for detecting and recognizing Arabic and English text in natural scene images. The detection pipeline combines a Swin Transformer Tiny backbone pretrained on ImageNet-22K, a Feature Pyramid Network for multi-scale fusion, and a DBNet++ head, achieving an F1-score of 94.52% on a strictly held-out 2,000-image ICDAR 2019 MLT test set. For recognition, PARSeq with permutation language modeling is trained on 656,868 cropped word samples (85/15 split), reaching 89.50% word accuracy and 94.13% character accuracy on an 865-character bilingual charset. The system includes a vertical-projection word-segmentation fallback and a smart RTL/LTR ordering algorithm. Ablations show FPN contributes +4.6 pp and DBNet++ +3.22 pp F1. Comparisons with prior baselines are reported as non-comparable references. The contribution is a fully reproducible bilingual Arabic-English pipeline.
English, Hindi and Marathi Multilingual optical character recognition (OCR) is challenging where the text is scanned on real-life images due to blur, uneven lighting, background noise, and mixed script writing. The paper introduces a lightweight script-aware OCR system that enhances recognition accuracy by using coordinated preprocessing, connected-component-based word localization, script routing, and confidence-based re-recognition instead of retraining the entire system. The pipeline consists of bilateral filtering, fast non-local means denoising, contrast-limited adaptive histogram equalization (CLAHE), adaptive thresholding, connected component analysis, script identification, and dual-pass Tesseract recognition. The algorithm is aimed at low-resource multilingual document comprehension in which interpretability, tunability, and relatively low computational efficiency are of significance. Experiments using a 5000 image corpus indicate a recognition accuracy of 93.0 on average and language-wise accuracies of 95% in English, 92% in Hindi and 90% in Marathi. The error rate in the characters within English, Hindi and Marathi were 4.2, 6.5 and 7.1 respectively and the script identification stage was 0.94 with a recall 0.93 and F1-score 0.935. The analysis of ablation reveals that adaptive thresholding and script identification are the most significant ones. The findings suggest that a well-designed, script-sensitive pipeline offers both a viable and technically rational solution to multilingual OCR of realistic images.
P. Patil, Monali Y.Khachane· 2026 International Conferenc...· 0 citations
An end-to-end assignment verification system that uses computer vision (CV) and natural language processing (NLP) to handle various complex types of modern academic documents. The solution uses multi-scale convolutional neural networks to capture writing and layout features. Using transformer-based structures for semantic parsing and segmentation. At its core is an adaptive cross-modal fusion mechanism, used to synchronize and merge visual and textual features to enhance the ability to recognize anomalous structures, fraudulent handwriting, and semantic mismatches. To train and test the system, we used a large, meticulously annotated dataset containing over 15,000 assignment papers, including handwritten, typed, and mixed-type papers. Many experiments compared the proposed framework with rule-based, vision-based, and language-based frameworks. The algorithm combinations performed excellently in document region detection, semantic mapping, and overall correctness. Cross-modal solutions can be used for real-time education. Further analysis showed the proportion of all subsystems and confirmed the necessity of reliably obtaining results from a wide range of heterogeneous and noisy data. There are limitations when dealing with charts that have different writing styles or do not follow standard drawing conventions, which may make understanding and grouping more difficult. Has it been proven that there is a reliable computational method to verify automatically high-fidelity tasks?
Yufei Ye, Wenyao Cao, Zerong Huang et al.· International Conference on...· 0 citations
Offline handwritten word recognition has progressed from whole-word classification to sequence transcription, yet many systems depend on large annotated corpora and exploit lexical regularities over explicit character evidence. This paper presents an alternative formulation for Arabic offline handwritten word recognition, treating characters as spatial objects detected via a Faster Region-Based Convolutional Neural Network rather than symbols generated by a one-dimensional decoder. We construct and release a character-level annotated subset of 2153 handwritten word images from a standard Arabic benchmark, exporting matched detection, sequence, and word-class labels. We also introduce an open-source subword exchange toolkit that creates a controlled structural-generalization benchmark by swapping subwords while preserving handwriting style. Experiments compare the proposed detector against whole-word and sequence-based baselines on both the original held-out split and the perturbed benchmark. Results show sequence models degrade sharply under structural recombination, whereas the proposed detector remains stable, achieving a 26.56% character error rate and 70.0% word accuracy on the perturbed benchmark. These findings demonstrate that explicit character localization provides a robust, data-efficient alternative for Arabic handwritten text recognition in low-resource settings.
Sofiane Medjram, Ruwaidah Saud Alnejaidi· Applied Sciences· 0 citations
A hybrid architecture that combines a Transformer-based DistilBERT model with a Graph Convolutional Network (GCN) that enhances detection by modeling structural relationships within text data is proposed.
The high rate of digitization of Arabic institutional archives has created masses of page images, which are hard to search where good textual metadata are unavailable. Lexical indexing can be supported by Optical Character Recognition (OCR), but its reliability suf ers from blur, skew, low contrast, compression artifacts, complex layout, and Arabic-specific script features like contextual letter forms, ligatures and dots, and optional diacritics. The paper will suggest and mathematically model a confidence-aware hybrid retrieval system in Arabic document images with the use of keyword queries. The framework integrates a lexical branch (OCR-based), a visual region-matching branch, branch-based score calibration, confidence gated fusion, re-ranking and keyword localization. The empirical part is limited by the measures which are directly justified by the measured datasets. On the FUNSD testing split (50 pages; 668 normalized single-token searches chosen in ground-truth annotations) an experiment of keyword retrieval at the page level was implemented. Using Tesseract 5.5.0 English OCR and BM25 indexing, OCR-BM25 achieved P@1 = 0.867, mAP = 0.678, NDCG@10 = 0.713, and MRR = 0.888; the confidence-weighted OCR variant achieved P@1 = 0.870, mAP = 0.672, NDCG@10 = 0.708, and MRR = 0.890. The given measured values confirm the auxiliary lexical retrieval protocol of non-Arabic noisy forms and demonstrate that raw OCR confidence is not uniformly beneficial when it comes to ranking metrics. They are not presented as Arabic hybrid retrieval performance. The assessed IFN/ENIT files include Arabic handwritten word images and segmentation XML but do not include page-level keyword relevance labels, whereas assessed RVL-CDIP test folders include document-class images instead of keyword annotations. It is based on this that the paper presents Arabic-oriented framework, reproducible evaluation protocol, measured auxiliary evidence and the identification of the dedicated Arabic retrieval benchmark that will be needed to complete target-domain validation.
A. S. Ibrahim· International Journal of App...· 0 citations
This paper presents a dual-architecture deep learning pipeline for real-time Arabic Sign Language (ArSL) recognition, designed to enhance communication accessibility for the Deaf and Hard of Hearing community. The current system recognises only isolated static Arabic letters; extending it to dynamic gestures and continuous signing is beyond the scope of this work and is identified as a key direction for future research. The proposed system integrates MediaPipe-based hand landmark detection with background removal and data augmentation, processing both static images and live video streams to translate isolated gestures into letters of the alphabet. We systematically compare two classification approaches under identical preprocessing and statistical evaluation protocols: (1) a fine-tuned ResNet18 achieving a mean accuracy of 97.8 ± 0.3% (95% CI [97.2%, 98.4%]), and (2) a hybrid architecture combining EfficientNet-B2 feature extraction with a Random Forest classifier, achieving 98.9 ± 0.2% (95% CI [98.5%, 99.3%]) on a diverse dataset of 7857 labelled RGB images from over 200 participants. This constitutes the first rigorous statistical comparison of these two paradigms specifically for ArSL recognition. Real-time deployment on consumer hardware achieves 19–22 FPS with per-frame latencies of 45–52 ms and no perceptible lag, demonstrating practical utility. Error analysis identifies anatomically similar letter pairs (Sad/Dad, Teh/Theh) as primary challenges. This work provides an accurate, efficient tool for real-world assistive technology applications.
Asmaa Y. Othman, Amira Gaber, S. M. El-Metwally· Discover Artificial Intellig...· 0 citations