{"ID":2850736,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.21590","arxiv_id":"2510.21590","title":"Restore Text First, Enhance Image Later: Two-Stage Scene Text Image Super-Resolution with Glyph Structure Guidance","abstract":"Current image super-resolution methods show strong performance on natural images but distort text, creating a fundamental trade-off between image quality and textual readability. To address this, we introduce TIGER (Text-Image Guided supEr-Resolution), a novel two-stage framework that breaks this trade-off through a \"text-first, image-later\" paradigm. TIGER explicitly decouples glyph restoration from image enhancement: it first reconstructs precise text structures and uses them to guide full-image super-resolution. This ensures high fidelity and readability. To support comprehensive training and evaluation, we present the UZ-ST (UltraZoom-Scene Text) dataset, the first Chinese scene text dataset with extreme zoom. Extensive experiments show TIGER achieves state-of-the-art performance, enhancing readability and image quality.","short_abstract":"Current image super-resolution methods show strong performance on natural images but distort text, creating a fundamental trade-off between image quality and textual readability. To address this, we introduce TIGER (Text-Image Guided supEr-Resolution), a novel two-stage framework that breaks this trade-off through a \"t...","url_abs":"https://arxiv.org/abs/2510.21590","url_pdf":"https://arxiv.org/pdf/2510.21590v2","authors":"[\"Minxing Luo\",\"Linlong Fan\",\"Wang Qiushi\",\"Ge Wu\",\"Yiyan Luo\",\"Yuhang Yu\",\"Jinwei Chen\",\"Yaxing Wang\",\"Qingnan Fan\",\"Jian Yang\"]","published":"2025-10-24T15:59:04Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[]","has_code":false}