SynthMap+ (Traditional Chinese) Synthetic Train Data for ICDAR'25 MapText Competition

Dataset of synthetic map images in traditional Chinese for the ICDAR'25 Competition on Historical Map Text Detection, Recognition, and Linking. Annotations and images follow the format described at the competition website. Please refer to [1] for the generation process and usage. We extend [1] to place text labels in horizontal and vertical writing directions. Train Annotations tc25synth_train.json Images train.zip Files tc25synth/train/*.jpg Tiles 45,000 Map Sheets - Words 296,348 Label Groups - Illegible Words 0 Truncated Words 0 Valid Words 296,348 [1] Lin, Y., & Chiang, Y. -Y. (2024). Hyper-local deformable transformers for text spotting on historical maps. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (pp. 5387-5397).