HarmonyText paper introduces mask-free scene text editing with reverse-to-refined synthesis
A new arXiv paper, HarmonyText, presents a natural language-guided, mask-free approach to editing text in real-world images, using reverse-to-refined pair synthesis and region-decoupled frequency learning.
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-11-16 05:00 · arXiv cs.CV
HarmonyText: Harmonious Scene Text Editing via Reverse-to-Refined Pair Synthesis and Region-Decoupled Frequency Learning