Authors: Yu Xie, Canhui Xu, Jielei Zhang, Pengyu Chen, Weihang Wang, Yuchen He, Peiyi Li, Yihan Meng, Longwen Gao

Affiliation: Bilibili Inc., QUST

Description: For the English MapText detection task, we employed DNTextSpotter, a novel denoising training method based on DeepSolo. For the Chinese MapText detection task, we utilized DeepSolo. Data augmentation techniques, including cropping, scaling, and adjustments to saturation and contrast, were applied. Pre-training was conducted using available real-world datasets such as TextOCR, TotalText, IC15, and MLT2017. Post-processing methods were also adopted.

Authors: Mengjie Zou, Tianhao Dai, Remi Petitpierre, Beatrice Vaienti, Frederic Kaplan, Isabella di Lenardo

Affiliation: EPFL, Swiss Federal Institute of Technology in Lausanne

Email: remi.petitpierre@epfl.ch

Description: For word detection and recognition, our approach relies on DeepSolo, whose architecture is derived from Detection Transformers (DETR). In short, DeepSolo extracts hierarchical visual features from map images and processes them through an encoder-decoder architecture to detect words as segments bounded by Bézier curves. The model specifically returns four control points of central Bézier curves per word and then uniformly samples query points along these curves to segment, classify, and delineate each text instance precisely. To resolve duplicate word detections, we implement a postprocessing step inspired by Non-Maximum Suppression. It involves calculating the Fréchet distance between the Bézier curves of potential duplicate word pairs, or "directional synonyms", and merging those below a defined threshold. More details on the model, algorithms, and specific implementation are provided in our separate article [1].

The model training leverages several real and synthetical datasets: ICDAR MapText [2], MapKuratorHuman [3], SynthMap [3], and Paris and Jerusalem Maps Text Dataset [4].

References:

[1] Zou, M., Dai, T., Petitpierre, R., Vaienti, B., Kaplan, F., & di Lenardo I. (2025). Recognizing and Sequencing Multi-word Texts in Maps Using an Attentive Pointer.

[2] Lin, Y., Li, Z., Chiang Y.Y., & Weinman J. (2024). Rumsey Train and Validation Data for ICDAR'24 MapText Competition (Version 1.3). Zenodo. https://doi.org/10.5281/zenodo.11516933

[3] Kim, J., Li, Z., Lin Y., Namgung, M., Jang, L., & Chiang Y.Y. (2023) The mapKurator System: A Complete Pipeline for Extracting and Linking Text from Historical Maps. In: Proceedings of the 31st ACM International Conference on Advances in Geographic Information Systems. https://arxiv.org/abs/2306.17059

[4] Dai, T., Johnson, K., Petitpierre, R., Vaienti, B., & di Lenardo, I. (2025). Paris and Jerusalem Maps Text Dataset (Version 1.0.0). Zenodo. https://doi.org/10.5281/zenodo.14982662

Authors: Yu Xie, Jielei Zhang, Ziyue Wang, Yuchen He, Yihan Meng, Weihang Wang, Peiyi Li, Longwen Gao, Qian Qiao

Affiliation: Bilibili Inc.

Description: In the detection task of MapText, we employed ViTAE-v2 to extract global features, utilizing an encoder-decoder network architecture (DeepSolo). Data augmentation techniques such as cropping, scaling, saturation, and contrast adjustment were applied. Pre-training was conducted using available real datasets (TextOCR, TotalText, IC15, MLT2017). Post-processing methods were also adopted.