A production-oriented PDF visual element parser achieves ≥96% detection accuracy and 93% caption association accuracy using heuristics and layout rules, outperforming prior parsers and vision-language models on benchmarks while cutting latency by more than 2×.
Learning transferable visual models from natural language supervision
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Lightweight and Production-Ready PDF Visual Element Parsing
A production-oriented PDF visual element parser achieves ≥96% detection accuracy and 93% caption association accuracy using heuristics and layout rules, outperforming prior parsers and vision-language models on benchmarks while cutting latency by more than 2×.