Enhancing Visual Localization with Cross-Domain Image Generation

Yuanze Wang, Yichao Yan, Shiming Song, Songchang Jin, Yilan Huang, Xingdong Sheng, Dianxi Shi
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:65075-65090, 2025.

Abstract

Visual localization aims to predict the absolute camera pose for a single query image. However, predominant methods focus on single-camera images and scenes with limited appearance variations, limiting their applicability to cross-domain scenes commonly encountered in real-world applications. Furthermore, the long-tail distribution of cross-domain datasets poses additional challenges for visual localization. In this work, we propose a novel cross-domain data generation method to enhance visual localization methods. To achieve this, we first construct a cross-domain 3DGS to accurately model photometric variations and mitigate the interference of dynamic objects in large-scale scenes. We introduce a text-guided image editing model to enhance data diversity for addressing the long-tail distribution problem and design an effective fine-tuning strategy for it. Then, we develop an anchor-based method to generate high-quality datasets for visual localization. Finally, we introduce positional attention to address data ambiguities in cross-camera images. Extensive experiments show that our method achieves state-of-the-art accuracy, outperforming existing cross-domain visual localization methods by an average of 59% across all domains. Project page: https://yzwang-sjtu.github.io/CDG-Loc.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-wang25dv, title = {Enhancing Visual Localization with Cross-Domain Image Generation}, author = {Wang, Yuanze and Yan, Yichao and Song, Shiming and Jin, Songchang and Huang, Yilan and Sheng, Xingdong and Shi, Dianxi}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {65075--65090}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/wang25dv/wang25dv.pdf}, url = {https://proceedings.mlr.press/v267/wang25dv.html}, abstract = {Visual localization aims to predict the absolute camera pose for a single query image. However, predominant methods focus on single-camera images and scenes with limited appearance variations, limiting their applicability to cross-domain scenes commonly encountered in real-world applications. Furthermore, the long-tail distribution of cross-domain datasets poses additional challenges for visual localization. In this work, we propose a novel cross-domain data generation method to enhance visual localization methods. To achieve this, we first construct a cross-domain 3DGS to accurately model photometric variations and mitigate the interference of dynamic objects in large-scale scenes. We introduce a text-guided image editing model to enhance data diversity for addressing the long-tail distribution problem and design an effective fine-tuning strategy for it. Then, we develop an anchor-based method to generate high-quality datasets for visual localization. Finally, we introduce positional attention to address data ambiguities in cross-camera images. Extensive experiments show that our method achieves state-of-the-art accuracy, outperforming existing cross-domain visual localization methods by an average of 59% across all domains. Project page: https://yzwang-sjtu.github.io/CDG-Loc.} }
Endnote
%0 Conference Paper %T Enhancing Visual Localization with Cross-Domain Image Generation %A Yuanze Wang %A Yichao Yan %A Shiming Song %A Songchang Jin %A Yilan Huang %A Xingdong Sheng %A Dianxi Shi %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-wang25dv %I PMLR %P 65075--65090 %U https://proceedings.mlr.press/v267/wang25dv.html %V 267 %X Visual localization aims to predict the absolute camera pose for a single query image. However, predominant methods focus on single-camera images and scenes with limited appearance variations, limiting their applicability to cross-domain scenes commonly encountered in real-world applications. Furthermore, the long-tail distribution of cross-domain datasets poses additional challenges for visual localization. In this work, we propose a novel cross-domain data generation method to enhance visual localization methods. To achieve this, we first construct a cross-domain 3DGS to accurately model photometric variations and mitigate the interference of dynamic objects in large-scale scenes. We introduce a text-guided image editing model to enhance data diversity for addressing the long-tail distribution problem and design an effective fine-tuning strategy for it. Then, we develop an anchor-based method to generate high-quality datasets for visual localization. Finally, we introduce positional attention to address data ambiguities in cross-camera images. Extensive experiments show that our method achieves state-of-the-art accuracy, outperforming existing cross-domain visual localization methods by an average of 59% across all domains. Project page: https://yzwang-sjtu.github.io/CDG-Loc.
APA
Wang, Y., Yan, Y., Song, S., Jin, S., Huang, Y., Sheng, X. & Shi, D.. (2025). Enhancing Visual Localization with Cross-Domain Image Generation. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:65075-65090 Available from https://proceedings.mlr.press/v267/wang25dv.html.

Related Material