method · first seen 1 Nov 2026
zero-shot optical character recognition with vision-language models
Also: zero-shot OCR
Using general vision-language models to transcribe text in images without task-specific training or fine-tuning.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Benchmarking Vision-Language Models for Urdu Nastaliq Handwriting RecognitionPapers · 1 Nov 2026 · Haseeb Yaqoob, Fatima Azfar, Emaan Arif Khan
Latest sources
- Benchmarking Vision-Language Models for Urdu Nastaliq Handwriting RecognitionPapers · 1 Nov 2026 · Haseeb Yaqoob, Fatima Azfar, Emaan Arif Khan