Urban scene segmentation is essential for 3D city modeling and plays a crucial role in various remote sensing applications, including urban planning and environmental monitoring. While integrating knowledge graphs with scene segmentation has improved accuracy, existing methods often depend on dataset-specific knowledge graphs, limiting their generalizability across diverse remote sensing data. To address this, we propose a novel framework that leverages large language models (LLMs) to construct a universal knowledge graph from multi-source geospatial data and incorporate it into remote sensing semantic segmentation tasks, enhancing adaptability and robustness in urban scene understanding. Specifically, the framework comprises two key components: (1) a Graph Construction module that employs LLMs to extract cross-domain semantic relationships and build a universal knowledge graph, and (2) a Knowledge Graph Fusion module (KGFusion) that incorporates the graph into a semantic segmentation network to enhance semantic understanding. To evaluate the adaptability of our method across diverse domains, we curated a mixed dataset encompassing urban, rural, and port scenes. Experimental findings validate the efficiency and adaptability of our method, achieving 70.94% mIoU on the UAVid dataset and 63.23% on the Mixed dataset, outperforming the baseline by 0.43% and 1.04%, respectively. These results validate the robustness of our method in cross-domain scenarios and highlight its potential for broader applications in complex urban environments.

Gaze estimation is a cornerstone of applications such as human–computer interaction and behavioral analysis, e.g., for intelligent transport systems. Nevertheless, existing methods predominantly rely on coarse-grained features from deep layers of visual encoders, overlooking the critical role that fine-grained details from shallow layers play in gaze estimation. To address this gap, we propose a novel Hierarchical Fine-Grained Attention Decoder (HFGAD), a lightweight fine-grained decoder that emphasizes the importance of shallow-layer information in gaze estimation. Specifically, HFGAD integrates a fine-grained amplifier MSCSA that employs multi-scale spatial-channel attention to direct focus toward gaze-relevant regions, and also incorporates a shallow-to-deep fusion module SFM to facilitate interaction between coarse-grained and fine-grained information. Extensive experiments on three benchmark datasets demonstrate the superiority of HFGAD over existing methods, achieving a remarkable 1.13° improvement in gaze estimation accuracy for in-car scenarios.

Remote sensing provides vast unlabeled imagery at low cost, yet annotation remains expensive, making self-supervised learning (SSL) well suited to this domain. However, existing DINO-style SSL frameworks are not well suited to remote sensing imagery, where object extents vary substantially and standard multi-crop view generation often introduces cross-scale inconsistency. This issue is particularly severe for small objects and elongated structures, whose discriminative features can be lost under scale transformations. To address this limitation, we propose DINO-MS (DINO with multi-scale consistency), a scale-consistent SSL framework for remote sensing imagery. The key idea is to construct feature-aligned cross-scale local views and explicitly enforce prediction-level agreement among them. Specifically, DINO-MS first adopts a co-located multi-scale cropping strategy to sample local views from the same spatial location at different crop scales, and then introduces a local consistency loss that works jointly with the original DINO local-to-global objective. Extensive experiments on land-use classification and change detection benchmarks show that DINO-MS generally improves downstream transfer performance. Notably, on EuroSAT, it improves per-class accuracy from 80.60% to 87.80% for Highway and from 88.00% to 91.60% for River with DINO-MC, confirming its advantage for categories dominated by small objects.