A Benchmark for Chinese-English Scene Text Image Super-resolution
This addresses a domain-specific problem for researchers in text image processing by providing a new benchmark and method for Chinese text super-resolution, though it is incremental as it builds on existing STISR work.
The paper tackles the lack of benchmarks for Chinese-English scene text image super-resolution by introducing the Real-CE dataset with 1,935/783 LR-HR pairs and 33,789 text lines, and proposes an edge-aware learning method that improves reconstruction of complex Chinese characters.
Scene Text Image Super-resolution (STISR) aims to recover high-resolution (HR) scene text images with visually pleasant and readable text content from the given low-resolution (LR) input. Most existing works focus on recovering English texts, which have relatively simple character structures, while little work has been done on the more challenging Chinese texts with diverse and complex character structures. In this paper, we propose a real-world Chinese-English benchmark dataset, namely Real-CE, for the task of STISR with the emphasis on restoring structurally complex Chinese characters. The benchmark provides 1,935/783 real-world LR-HR text image pairs~(contains 33,789 text lines in total) for training/testing in 2$\times$ and 4$\times$ zooming modes, complemented by detailed annotations, including detection boxes and text transcripts. Moreover, we design an edge-aware learning method, which provides structural supervision in image and feature domains, to effectively reconstruct the dense structures of Chinese characters. We conduct experiments on the proposed Real-CE benchmark and evaluate the existing STISR models with and without our edge-aware loss. The benchmark, including data and source code, is available at https://github.com/mjq11302010044/Real-CE.