In this paper, we propose a large language models (LLMs) enhanced path planning framework for a cellular-connected unmanned aerial vehicle (UAV), which can significantly improve the efficiency of path planning without sacrificing path quality. The ray tracing model is used to generate the radio map of UAV cellular connectivity from real-world geospatial data. By introducing quadtree-assisted representation, the radio map is formatted into prompts that can be understood by LLMs, bridging the gap between simplified obstacle environments and realistic electromagnetic environments. Meanwhile, a combined approach using LLMs and the graph-based search method is employed to find a path which can guarantee reliable connections between the UAV and its associated ground base station throughout the flight. Simulation results demonstrate that our proposed LLM-enhanced framework not only finds the path that guarantees communication quality, but also addresses computational and memory limitations of the conventional method, with a significant reduction in operations and storage.

While significant progress has been made in large-scale scene representation using Neural Radiance Fields (NeRF), several limitations remain. For instance, most methods still rely on the original coarse-to-fine sampling strategy, leading to an inefficient rendering process. Additionally, to model larger scenes, these methods often use complex network models, resulting in redundant model parameters. To address these issues, we propose a novel model with adaptive sampling and feature-aware compression for large-scale scene rendering, named ASFC-NeRF. We first introduce a weight prediction network to replace the original coarse sampling strategy, then employ a teacher network and depth constraints for knowledge distillation in the early stages of training to enhance the high-fidelity of the scene. Furthermore, we optimize the number of Grids and the channels of Planes and prune the network to efficiently compress model parameters. Experimental results demonstrate that our method significantly accelerates the rendering process and greatly reduces parameter quantity while maintaining or only slightly lowering image quality. Therefore, ASFC-NeRF exhibits advantages in comprehensive performance and practicality.

We introduce NeRF-GS, a novel framework that jointly optimizes Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS). This framework leverages the inherent continuous spatial representation of NeRF to mitigate several limitations of 3DGS, including sensitivity to Gaussian initialization, limited spatial awareness, and weak inter-Gaussian correlations, thereby enhancing its performance. In NeRF-GS, we revisit the design of 3DGS and progressively align its spatial features with NeRF, enabling both representations to be optimized within the same scene through shared 3D spatial information. We further address the formal distinctions between the two approaches by optimizing residual vectors for both implicit features and Gaussian positions to enhance the personalized capabilities of 3DGS. Experimental results on benchmark datasets show that NeRF-GS surpasses existing methods and achieves state-of-the-art performance. This outcome confirms that NeRF and 3DGS are complementary rather than competing, offering new insights into hybrid approaches that combine 3DGS and NeRF for efficient 3D scene representation.

Knowledge distillation (KD) stands as an efficient technique for compressing models, typically employing a teacher-student framework. Nevertheless, optimizing KD to yield models with reduced parameters and enhanced performance remains an area warranting deeper investigation. In this article, we recognize the significance of preserving structural consistency to enhance knowledge transfer efficiency between networks. Leveraging this insight, we introduce a novel approach termed Student-Driven Knowledge Distillation (SDKD), which integrates a proxy teacher intermediary between the primary teacher and student model. Specifically, we construct the architecture of the proxy teacher entirely based on the student network to generate logits that closely align with the distribution of the student network. Besides, we propose a Feature Fusion Block (FFB) to integrate features from the teacher network into the proxy teacher. FFB can not only provide high-quality feature-based knowledge for distillation but also impart response-based knowledge to facilitate the learning process. Extensive experiments illustrate that SDKD outperforms 29 state-of-the-art methods on several tasks, including image classification, semantic segmentation, and depth estimation.

We propose the Semantic-aware NeRF Experts (SANE), which fully exploits the intrinsic characteristics of large-scale scenes, including semantics and material features, to achieve high-quality novel view synthesis results and provide accurate 3D semantic information. SANE begins by building a semantic Mixture of Experts (MoE), utilizing a learnable gating network to semantically partition the scene into blocks for corresponding NeRF experts. We then develop a semantic volume rendering scheme that integrates discrete semantics into the end-to-end differentiable process of NeRF, enabling refined semantic labeling of each scene point. Additionally, we implement a dual-implicit encoding strategy: intra-block encoding captures lighting variations across viewpoints, while inter-block one captures texture features among different semantic objects. Experiments on benchmark datasets show that SANE delivers higher-quality scene representations and effective semantic decomposition for downstream tasks, such as precise editing of large-scale scenes based on semantics.

VR applications rely on the user's ability to explore the virtual scene efficiently. In complex scenes, occlusions limit what the user can see from a given location, and the user has to navigate the viewpoint around occluders to gain line of sight to the hidden parts of the scene. When the disoccluded regions prove to be of no interest, the user has to retrace their path, making scene exploration inefficient. Furthermore, the user might not be able to assume a viewpoint that would reveal the occluded regions due to physical limitations, such as obstacles in the real world hosting the VR application, viewpoints beyond the tracked area, or viewpoints above the user's head that cannot be reached by walking. Several occlusion management methods have been proposed in visualization research, such as top view, X-ray, and multiperspective visualization, which help the user see more from the current position, having the potential to improve the exploration efficiency of complex scenes. This paper reports on a study that investigates the potential of these three occlusion management methods in the context of VR applications, compared to conventional navigation. Participants were required to explore two virtual scenes to purchase five items in a virtual Supermarket, and to find three people in a virtual parking garage. The task performance metrics were task completion time, total distance traveled, and total head rotation. The study also measured user spatial awareness, depth perception, and simulator sickness. The results indicate that users benefit from top view visualization which helps them learn the scene layout and helps them understand their position within the scene, but the top view does not let the user find targets easily due to occlusions in the vertical direction, and due to the small image footprint of the targets. The X-ray visualization method worked better in the garage scene, a scene with a few big occluders and a low occlusion depth complexity' and less well in the Supermarket scene, a scene with many small occluders that create high occlusion depth complexity. The multi-perspective visualization method achieves better performance than the top view method and the X-ray method, in both scenes. There are no significant differences between the three methods and the conventional method in terms of spatial awareness, depth perception, and simulator sickness.