A new 1,477-frame video OCR benchmark shows GPT-4o and Gemini-1.5 Pro beat EasyOCR and RapidOCR on average accuracy, but the evaluation is preliminary and lacks statistical rigor.
Gemini 1.5 Pro: Pushing the boundaries of multimodal learning, 2024
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Benchmarking Vision-Language Models on Optical Character Recognition in Dynamic Video Environments
A new 1,477-frame video OCR benchmark shows GPT-4o and Gemini-1.5 Pro beat EasyOCR and RapidOCR on average accuracy, but the evaluation is preliminary and lacks statistical rigor.