conftrace_

numerical fidelity

1 paper

Explore in graph

Co-occurring keywords · papers mentioning both

dual-head image encoder (1) fine details (1) vision-language model (1) structural cues (1) preference learning (1)

Papers

Twin-T & TwintVQA: A Reliable Structure-Detail Separating VLM and a Comprehensive Benchmark for Chart and Table Tasks CVPR 2026