Large Language Models (LLMs) have shown immense potential in Knowledge Graph Completion (KGC), yet bridging the modality gap between continuous graph embeddings and discrete LLM tokens remains a critical challenge.While recent quantization-based approaches attempt to align these modalities, they typically treat quantization as flat numerical compression, resulting in semantically entangled codes that fail to mirror the hierarchical nature of human reasoning.In this paper, we propose GS-Quant, a novel framework that generates semantically coherent and structurally stratified discrete codes for KG entities.Unlike prior methods, GS-Quant is grounded in the insight that entity representations should follow a linguistic coarse-to-fine logic.We introduce a Granular Semantic Enhancement module that injects hierarchical knowledge into the codebook, ensuring that earlier codes capture global semantic categories while later codes refine specific attributes.Furthermore, a Generative Structural Reconstruction module imposes causal dependencies on the code sequence, transforming independent discrete units into structured semantic descriptors.By expanding the LLM vocabulary with these learned codes, we enable the model to reason over graph structures isomorphically to natural language generation.Experimental results demonstrate that GS-Quant significantly outperforms existing text-based and embedding-based baselines.Our code is publicly available at https: //github.com/mikumifa/GS-Quant.

Graph anomaly detection (GAD) aims to identify nodes that deviate from normal patterns in structure or features. While recent GNN-based approaches have advanced this task, they struggle with two major challenges: 1) homophily disparity, where nodes exhibit varying homophily at both class and node levels; and 2) limited scalability, as many methods rely on costly whole-graph operations. To address them, we propose SAGAD, a Scalable and Adaptive framework for GAD. SAGAD precomputes multi-hop embeddings and applies reparameterized Chebyshev filters to extract low- and high-frequency information, enabling efficient training and capturing both homophilic and heterophilic patterns. To mitigate node-level homophily disparity, we introduce an Anomaly Context-Aware Adaptive Fusion, which adaptively fuses low- and high-pass embeddings using fusion coefficients conditioned on Rayleigh Quotient-guided anomalous subgraph structures for each node. To alleviate class-level disparity, we design a Frequency Preference Guidance Loss, which encourages anomalies to preserve more high-frequency information than normal nodes. SAGAD supports mini-batch training, achieves linear time and space complexity, and drastically reduces memory usage on large-scale graphs. Theoretically, SAGAD ensures asymptotic linear separability between normal and abnormal nodes under mild conditions. Extensive experiments on 10 benchmarks confirm SAGAD's superior accuracy and scalability over state-of-the-art methods.

Video-based Dynamic Mesh Coding (V-DMC), being standardized by the MPEG-3DGH group, seeks to establish an efficient standard for compressing dynamic meshes with time-varying vertex positions, connectivity, and attributes. By adopting a subdivision and video-based framework, V-DMC effectively leverages both advanced static mesh and video codecs, achieving state-of-the-art dynamic mesh coding efficiency. However, in V-DMC, inter-frame coding—one of the key coding tools—is applied only to a limited subset of frames and remains computationally expensive. To overcome this limitation, we propose a novel inter-frame coding framework for V-DMC that extends applicability to a much larger portion of frames while achieving both high efficiency and low complexity. Specifically, our method consists of four modules: (a) supervoxel-based shape matching for robust and efficient motion estimation; (b) embedded graph deformation for accurate geometry tracking; (c) early interframe coding mode decision for accelerated rate-distortion optimization; and (d) UV atlas tracking for generating temporally consistent texture images. Experimental results on the MPEG dynamic mesh dataset show that the proposed method achieves BD-rate improvements of −10.1%, −10.1%, −14.7%, −17.6%, and −15.5% for D1, D2, Luma, Cb, and Cr, respectively, along with a 5% decrease in encoding runtime compared to V-DMC reference software.

The MPEG-3DGH group is developing a new standard for dynamic mesh encoding, known as Video-based Dynamic Mesh Coding (V-DMC). One of the main challenges in the V-DMC encoding process is extracting temporally corresponding information required for the inter-frame coding when dealing with inputs that have time-varying connectivity. To provide an efficient and effective solution, we propose a fast supervoxel-based base mesh generation method complient with V-DMC inter-frame coding. Our method includes four steps: 1) Initial supervoxel segmentation on two given frames; 2) Supervoxel-based shape matching to extract an initial correspondence set; 3) Refined supervoxel segmentation and shape matching with the assistance of the initial correspondence set, which outputs a refined correspondence set with more correspondence vertices and higher accuracy; 4) Generation of inter-frame base meshes using the refined correspondence set. Experiments on the MPEG V-DMC test sequences demonstrate that our method achieves BD-rate improvements over the VDMC v6.0 in terms of D1, D2, Luma, Cb, and Cr metrics by -9.2%, -9.3%, -14.1%, -14.6%, and -9.7%, respectively. Additionally, the proposed method only increases the encoding time from 100% to 105% while suppressing the decoding time from 100% to 99%.

The MPEG-3DGH group is developing a new standard for dynamic mesh encoding, known as Video-based Dynamic Mesh Coding (V-DMC). While V-DMC leverages a base mesh framework to exploit temporal correlations in the geometry, it does not account for extracting temporal information during UV atlas generation. To extend the efficiency of UVAtlas from single-frame to multi-frame scenarios, we propose a temporally consistent UVAtlas generation method compatible with V-DMC. Our method begins by applying keypoint tracking to capture the motion of the input sequence. Next, we perform patch partitioning and mapping to ensure consistent patches across frames. Each patch is then parameterized by minimizing geodesic distance distortion. Finally, all matched patches are packed into the same position within the texture domain. Our approach successfully generates consistent texture maps across an entire GOF (Group of Frames), significantly reducing the texture bitrate. Experimental results on the MPEG V-DMC test sequences demonstrate that our method achieves significant BD-rate improvements over the anchor of −13.9%, −13.9%, −10.6%, −14.5%, and −12.0% for D1, D2, Luma, Cb, and Cr PSNR metrics, respectively.

The Video-based Dynamic Mesh Coding (V-DMC), currently under development by the MPEG-3DGH group, employs a simplification and approximation framework for geometry coding. First, the input mesh is simplified into a base mesh with fewer faces, edges, and vertices. The base mesh is then subdivided and deformed to approximate the original mesh, with the offsets between the subdivided mesh and the deformed mesh referred to as displacements. Optimizing both the base mesh and the deformed mesh is crucial to achieving high coding quality. This process is termed geometry parametrization in V-DMC. In this paper, we propose a stabilization method for geometry parametrization in V-DMC to ensure robust results. Experimental results on the MPEG VDMC test sequences show that our method not only enhances the numerical stability of the geometry parametrization process but also improves quantitative coding performance and visual quality. Specifically, our method achieves BD-rate improvements over the anchor of -0.2%, -0.2%, -0.2%, -0.2%, and -0.2% for D1, D2, Luma, Cb, and Cr PSNR metrics, respectively. The proposed method has been adopted and integrated into the MPEG V-DMC reference software.

Children around 5 years old exhibit a strong inclination to help others. However, they show less helping behavior when others are present, which is known as the bystander effect. As social robots enter children’s lives, children’s helping behavior may also be affected by social robots. Therefore, the purpose of this study is to find out whether children also exhibit bystander effects when confronted with a distressed robot, and its underlying factors contributing to this phenomenon. Sixty participants (Mage = 66.10 months, SD = 4.46) were randomly assigned to three conditions: alone condition, bystander condition and bystander-unavailable condition. Children’s helping behavior was observed while interacting with the robot recipient, and then they participated in an animacy interview. The results revealed that children can attribute animacy to the robot, however, in bystander condition, their helping behavior was significantly less than in other conditions when confronted with a robot recipient, mainly due to the diffusion of responsibility.

The Video-based Dynamic Mesh Coding (V-DMC) standard exploits the temporal correlation by tracing the motions of the vertices, which applies only to tracked frames with one-to-one vertex correspondence. For non-tracked frames, only intra mode is applied. This paper proposes an embedded graph representation method that can efficiently represent inter-frame differences for both tracked and non-tracked frames. First, we construct an embedded graph by simplifying the given mesh. Then, we compute a set of affine transformations on graph nodes and use their linear combinations to represent the inter-frame difference. Finally, we apply our implementation to increase the number of predicted frames (P-frames) and thus improve the resulting coding performance. Evaluations on Moving Picture Experts Group (MPEG) test sequences demonstrate the significant rate distortion improvements achieved by our method over V-DMC. The proposed method is highly compliant with V-DMC, and a part of it has been adopted into the V-DMC reference software during the MPEG-3DGH 144th meeting.

The MPEG-3DGH group is advancing a new standard for dynamic mesh encoding, known as Video-based Dynamic Mesh Coding (V-DMC). While V-DMC improves efficiency by toggling between intra- and inter-frame coding, it struggles with dynamic meshes that change inconsistently across frames. To address this issue, we introduce a novel partial inter-frame coding mode within the V-DMC framework. The method initiates by partitioning the adjacent frames into temporally aligned segments. These segments are then classified as suitable for either intra-frame or inter-frame coding based on the detection of inconsistent changes. Leveraging these segmentations, we introduce a partial inter-frame coding mode that allows for the selective application of intra- and inter-frame coding at the level of individual segments. Experiments on the MPEG V-DMC test sequences demonstrate that our method can achieve BD-rate improvement over the anchor in terms of D1, D2, Luma, Cb, and Cr PSNRs by -2.1%,-2.7%,-4.7%,-5.4%, and -4.5%, respectively.

The inter-frame coding of dynamic meshes with time varying topology is still under development in the current Video-based Dynamic Mesh Coding (V-DMC) standard. To address this issue and improve the coding efficiency, we propose a temporally-consistent re-meshing method. In particular, we introduce a robust inter-surface mapping framework to re-mesh the input meshes so that they have one-to-one vertex and face correspondence. Then, we simultaneously decimate the re-meshed input meshes to generate temporally-consistent base meshes, which is a key requirement for applying inter-frame coding in V-DMC. The evaluations on Moving Picture Experts Group (MPEG) test sequences demonstrate that our method can achieve rate-distortion performance superior to that of V-DMC.

Generating hyperspectral point cloud from hyperspectral images (HSIs) and light detection and ranging (LiDAR) has become more and more common in the remote sensing field and supported various applications. One challenge here is that hyperspectral imaging is a passive imaging method and is suffering from shadows in a natural scene. Intrinsic information recovery can effectively eliminate the spectral variation caused by illumination changes; however, it assumes a uniform light and neglects the shadows in the scene. In this article, we provide a novel hyperspectral point cloud intrinsic model that can detect the shaded regions and recover reflectance information in them. We first estimate the global illumination of the scene using an intrinsic information recovery method. Then, we perform supervoxel segmentation on hyperspectral point cloud to calculate the blocking relation of supervoxels and therefore accurately detect shaded regions. Finally, we estimate the illumination and reflectance of shaded regions based on an illumination-invariant spectral prior. The experimental results show that the proposed method can effectively detect shaded areas and robustly generate shadow-less intrinsic hyperspectral point cloud.

Due to interference with remote imaging by some natural factors, the multitemporal analysis ability is limited by the spectral drift between images. In this article, a new approach to optimize the existing multitemporal analysis system is proposed: multitemporal intrinsic image decomposition (MIID). The MIID method is designed to extract common spectral reflectance from multitemporal images. With MIID, multitemporal classification, changing detection, and index extracting will become extremely easy and more accurate. Firstly, without considering land cover change, the general MIID framework is proposed by adding local temporal–spatial energy constraints in traditional intrinsic images decomposition. On this basis, an improved MIID method with change detection (CD) (CD-MIID) capability is proposed to make the model adapt to the land cover change situation. Finally, specific steps of how to use MIID methods in the multitemporal analysis are given. Multitemporal multispectral/hyperspectral remote sensing images from GF-1, GF-2, GF-5, Landsat TM, and two groups of captured datasets with reflectance truth map are used to evaluate the performance. The experimental results show the following two points: first, the MIID methods achieve better extraction results of spectral reflectance. Second, the proposed MIID methods have better performance both on multitemporal classification and CD.

Hyperspectral intrinsic image decomposition (HyperIID) has been proven to be a very useful approach to reduce the spectral uncertainty in the remote sensing imaging process and improve the classification. In this article, a new HyperIID with enhanced spatial information, called ESI-IID, is proposed to overcome the deficiency of low spatial resolution in the existing HyperIID methods. With the aid of high-resolution (HR) panchromatic (PAN) image, the proposed method embeds the HR spatial information into the intrinsic decomposition model and enhances spatial details of the intrinsic component. The proposed ESI-IID introduces three constraints: 1) we make the constraint on spectral information to protect it from distortion during the spatial resolution enhancement process; 2) we add the constraint on spatial information to make sure that the details of edges will be well kept; and 3) based on the assumption that the reflectance component has a strong correlation in the local neighborhood, we add the self-constraint on reflectance component, in which the similarity matrix consists of two parts extracted from hyperspectral images and PAN image, respectively. Finally, we build a matrix energy function according to the aforementioned constraints and solve it by finding the minimum Frobenius norm iteratively. Both visual and quantitative experiments on simulated and real datasets demonstrate that the proposed method outperforms other alternative methods with high reliability.

The combination of spectral and 3-D elevation information provided by hyperspectral images (HSIs) and Light Detection and Ranging (LiDAR) has gained increased attention in the remote sensing field and enabled numerous applications. While various methods have been proposed to fuse these two data streams in pixel, feature, or decision level, a deeper view into the intrinsic relation of surface geometry, material reflectance, and environment illumination is still lacking. In this article, we present a novel supervoxel-based joint intrinsic decomposition framework for HSIs and LiDAR. First, we proposed a novel intrinsic scene model for HSIs and LiDAR point cloud, which tells how we can map LiDAR point cloud into HSI pixels with point-cloud-level normals, reflectance, and incident light direction. Then, we extract supervoxels from the LiDAR point cloud using a graph-based supervoxel method. Finally, we formulate the intrinsic decomposition problem within a supervoxel-based framework which can be optimized effectively and efficiently. The outputs of the proposed model are intrinsic scene properties like incident light direction and point-cloud-level hyperspectral reflectance, with which we can then generate intrinsic hyperspectral point cloud (IHSPC) where each point possesses not only 3-D coordinates and normals but also the reflectance over each wavelength. The performance of our approach is demonstrated with both synthetic and real data.

Intrinsic hyperspectral image decomposition (IHID) aims to recover physical scene properties such as reflectance and illumination from a given hyperspectral image (HSI), which directly respects the physical imaging process and can benefit many HSI processing tasks. It is a severely ill-posed problem and is challenging to solve using HSI alone. Additional geometric information provided by digital surface models (DSMs) can otherwise help immensely. While intrinsic image decomposition for RGB images and RGB-D images has been studied extensively during the past few decades and has seen significant progress, studies of the problem for other types of data, such as HSIs and DSMs, are still needed. It is much more challenging to handle an HSI with hundreds of channels than an RGB image with only three channels. Moreover, compared with RGB-D data, HSIs and DSM data usually have much lower spatial resolutions and more complicated land covers, making it difficult to extend the RGB-D intrinsic image method directly. In this article, we present a novel IHID framework for HSIs with DSM cues. Utilizing spherical-harmonic illumination, we first propose a convenient HSI rendering model with DSM, which describes the interplay of material reflectance, geometric distribution, and environment illumination. Then, we introduce local and nonlocal priors on reflectance that ensure the local smooth and global consistency of recovered reflectance. Experiments on synthetic and real data demonstrate that the proposed method outperforms the state-of-the-art methods and is robust to illumination changes.

Satellite videos provide rich dynamic information of observed scenes at a large spatial and temporal scale and will play an important role in the future space information network. This work devotes to revealing events of interest (EOI) from satellite video scenes by using a two-stream method. In satellite videos, individual frames reflect the static information like the basic scenes where the event was happening, while a sequence of frames determines the motion information. Considering these facts, a novel two-stream EOI detection framework is proposed, where one stream extracts static spatial information of satellite videos by AlexNet, whereas the other stream extracts the motion information using a local trajectories analysis method. First, the whole video scene is segmented into small spatial-temporal patches, where labeling EOI and non-EOI is completed. Next, the trajectories are extracted from 3-D satellite video cubes that are generated from event scene patches. Finally, this trajectory classification process is treated as a weak supervision learning problem and solved by sparse dictionary learning. The experimental results demonstrate that the proposed two-stream method is effective for EOI detection and has a huge potential for satellite video scenes analysis and understanding. The proposed method also outperforms the existing competitive models for video analysis.

Binary-only bug search has already drawn a lot attentions recently, due to the increasing growth of security breaches. Most of existing work focuses on searching by checking the similarity of code snippets. It is further required to check whether the function is patched or not. Unfortunately, this is still a manual effort for all existing code search based approaches. In this paper, we propose a novel approach for automatic patch detection. we build a patch detector by learning the feature representation from the patched code in the binary format. We utilize the feature encoding technique to make the binary code trainable, and build our neural network model to learn the patch feature for increasing detection accuracy. We have implemented a prototype called PATCHDETECTOR, and systematically evaluated its performance in terms of the accuracy and efficiency by using 1,600 OpenSSL binaries of 216,000 functions. Experimental results have shown that PATCHDETECTOR can effectively detect whether the target binary function is patched or not with the detection accuracy of 92% on average.

In this paper, a novel reflectance model is proposed to recover intrinsic images from remote sensing hyperspectral images (HSIs). Intrinsic image recovery is a well-known challenging and underconstrained problem in computer vision, and it becomes even more severely illposed for HSIs. To reduce the uncertainties and improve the recovery accuracy, two kinds of priors are introduced: 1) shading prior which describes the geometric relation between illuminate and object surface and 2) reflectance prior based on L1-graph coding, which describes the relation between pigment density with reflectance. These priors can effectively eliminate the reflectance inhomogeneity caused by surface normal changes or pigment density variations other than material changes. Then, a noniterative optimization method is proposed to combine the shading prior and reflectance prior, with which closed-form solutions can be derived and thus avoided falling into local optimums. The experimental results demonstrate that the proposed method can efficiently improve the spectral reflectance homogeneity within a class while preserving the image boundaries; it also produces a competitive performance with the state of the art when utilizing the extracted intrinsic hyperspectral reflectance feature in the task of HSI classification.

Intrinsic image decomposition (IID) of hyperspectral images (HSIs) aims to separate the reflectance cube and shading component from the original image data. The reflectance cube contains the spectral information reflecting the intrinsic properties of the material, whereas the shading component contains the spatial information reflecting geometric structure of the object like the surface orientation changes. From the perspective of hyperspectral image classification, combining spectral information with spatial information can be useful for improving the classification performance. In this paper, a new optimization algorithm is proposed for intrinsic image decomposition of hyperspectral images, and composite kernel learning (CKL) method is further utilized to combine reflectance with shading component.

In this paper, we propose a novel superpixel-based intrinsic image decomposition (SIID) framework for hyperspectral images. Intrinsic images are usually referred to the separation of shading and reflectance components from an input image. Considering the high dimensionality of hyperspectral images, we further decompose the shading component into the product of environment illumination and surface orientation changes, thus modeling the problem more properly. The proposed method consists of the following steps. First, we build two superpixel segmentation maps of different scales, i.e., a finer one that is oversegmented and a coarser one that is undersegmented. Based on the observation that the finer superpixel map achieves a higher segmentation accuracy, whereas the coarser superpixel map tends to reserve the objectness of the original image, we model the SIID decomposition problem in a matrix form based on the finer superpixel map and define a constraint matrix by integrating the information in the coarser superpixel map. The constraint matrix is introduced as a secondary constraint in order to make the ill-posed IID problem solvable. Finally, we transform the original decomposition problem into minimizing the Frobenius norm of the proposed matrix energy function and iteratively derive the solution. Our experimental results demonstrate that the proposed method is able to achieve a performance outperforming the state-of-the-art while making a great improvement in efficiency.