Search papers, labs, and topics across Lattice.
Visual Place Recognition (VPR) systems often fail under extreme weather, lighting, and occlusion shifts because existing training datasets lack appearance diversity for identical geographic locations. To resolve this, AdaptVPR uses a vision-language model and a route-aware generative pipeline to synthesize same-place hard positives across global appearance, local occlusion, and dual-perturbation routes, filtering candidates via geometric consistency and appearance diversity checks. Trained on the resulting 160K-image AdaptCities dataset, standard VPR baselines and vision foundation models achieve robust cross-domain retrieval, driving R@1 improvements of up to 9.2%.
Visual place recognition models often crumble under weather and lighting shifts, but injecting 160K geometrically verified, route-aware synthetic hard positives boosts R@1 retrieval by up to 9.2% across foundation backbones.
Visual Place Recognition (VPR) localizes a query image by retrieving database images of the same or nearby place, yet its robustness is often degraded by domain shifts arising from illumination, weather, seasonal changes, and dynamic occlusions. One contributing factor is the limited appearance diversity of the same place in existing training data. To address this issue, we propose AdaptVPR, a route-aware generative augmentation framework that constructs same-place hard positives for robust VPR training. AdaptVPR first uses a vision language model to parse scene attributes and estimate editing feasibility, while a rule-based scheduler determines the generation route according to editability scores and risk constraints. The generation process is decomposed into three complementary routes: the Global Appearance Route introduces global scene changes in weather, illumination, and time of day; the Local Occlusion Route inserts plausible dynamic occluders; and the Dual Route combines both types of perturbations to produce more challenging appearance shifts. Each generated candidate is evaluated using a VPR-oriented verification scheme based on geometric consistency and appearance diversity, reducing the risk of structural drift while ensuring sufficient appearance variation. Global candidates are generated once and rejected if verification fails, while Local Occlusion and Dual candidates use verification feedback for limited prompt refinement and regeneration. Using this framework, we construct AdaptCities, containing 160K verified synthetic same-place hard positives. Experiments across multiple VPR baselines and vision foundation backbones show consistent gains on standard benchmarks and substantial improvements under challenging domain shifts, with R@1 gains of up to 9.2%. The source code and data resources are publicly available at https://github.com/chenshunpeng/AdaptVPR.
