Get Started
Home
Topics
Search
Library
Research questionHow can scanned multilingual dictionaries be digitized into structured entries without losing characters or lexicographic structure?Scanned dictionaries often combine language-specific writing systems with dense layouts, abbreviations, and cross-references. A usable digital version must preserve both the text and the relationships that define each lexicographic entry.
AI
Computer Vision
Evaluation & Benchmarks
Multimodal Models
Natural Language Processing
Research Paper
Latest papersRecent research connected to this question, newest first.MUDIDI: A Two-Stage Framework for Multilingual Dictionary Digitization with Language ModelsThe evidence comes from MUDIDI, a two-stage digitization framework evaluated on human-annotated entries from 30 public-domain dictionaries spanning diverse writing systems, language families, and formats. It compares OCR systems, general-purpose LLMs, and vision-language models for character and markup preservation, then entry segmentation and mapping to SIL’s Multi-Dictionary Formatter schema; the reported results favor LLMs across most tested settings and indicate that dictionary introductions can improve digitization quality.research paper · Sep 2, 2026
Related questions
How can SAR-to-optical translation preserve task-relevant structure for downstream segmentation?How can pixel-text encoders robustly represent grounded multilingual text across resolutions under tight visual-token budgets?How can video-based sign dictionary retrieval generalize across signers and unseen languages?How can OCR diagnose and correct case-level errors while preserving rendering-equivalent outputs?