Human pose and shape (HPS) estimation presents challenges in diverse scenarios such as crowded scenes, person-person interactions, and single-view reconstruction. Existing approaches lack mechanisms to incorporate auxiliary "side information" that could enhance reconstruction accuracy in such challenging scenarios. Furthermore, the most accurate methods rely on cropped person detections and cannot exploit scene context while methods that process the whole image often fail to detect people and are less accurate than methods that use crops. While recent language-based methods explore HPS reasoning through large language or vision-language models, their metric accuracy is well below the state of the art. In contrast, we present PromptHMR, a transformer-based promptable method that reformulates HPS estimation through spatial and semantic prompts. Our method processes full images to maintain scene context and accepts multiple input modalities: spatial prompts like bounding boxes and masks, and semantic prompts like language descriptions or interaction labels. PromptHMR demonstrates robust performance across challenging scenarios: estimating people from bounding boxes as small as faces in crowded scenes, improving body shape estimation through language descriptions, modeling person-person interactions, and producing temporally coherent motions in videos. Experiments on benchmarks show that PromptHMR achieves state-of-the-art performance while offering flexible prompt-based control over the HPS estimation process.

Reconstructing physically plausible human motion from monocular videos remains a challenging problem in computer vision and graphics. Existing methods primarily focus on kinematics-based pose estimation, often leading to unrealistic results due to the lack of physical constraints. To address such artifacts, prior methods have typically relied on physics-based post-processing following the initial kinematics-based motion estimation. However, this two-stage design introduces error accumulation, ultimately limiting the overall reconstruction quality. In this paper, we present PhysHMR, a unified framework that directly learns a visual-to-action policy for humanoid control in a physics-based simulator, enabling motion reconstruction that is both physically grounded and visually aligned with the input video. A key component of our approach is the pixel-as-ray strategy, which lifts 2D keypoints into 3D spatial rays and transforms them into global space. These rays are incorporated as policy inputs, providing robust global pose guidance without depending on noisy 3D root predictions. This soft global grounding, combined with local visual features from a pretrained encoder, allows the policy to reason over both detailed pose and global positioning. To overcome the sample inefficiency of reinforcement learning, we further introduce a distillation scheme that transfers motion knowledge from a mocap-trained expert to the vision-conditioned policy, which is then refined using physically motivated reinforcement learning rewards. Extensive experiments demonstrate that PhysHMR produces high-fidelity, physically plausible motion across diverse scenarios, outperforming prior approaches in both visual accuracy and physical realism.

We introduce Gaussian Articulated Template Model (GART), an explicit, efficient, and expressive representation for non-rigid articulated subject capturing and rendering from monocular videos. GART utilizes a mixture of moving 3D Gaussians to explicitly approximate a deformable subject's geometry and appearance. It takes advantage of a categorical template model prior (SMPL, SMAL, etc.) with learnable forward skinning while further generalizing to more complex non-rigid deformations with novel latent bones. GART can be reconstructed via differentiable rendering from monocular videos in seconds or minutes and rendered in novel poses faster than 150fps.

The rapid adoption of IPv6 has increased network access scale while also escalating the threat of Distributed Denial of Service (DDoS) attacks. By the time a DDoS attack is recognized, the overwhelming volume of attack traffic has already made mitigation extremely difficult. Therefore, continuous network monitoring is essential for early warning and defense preparation against DDoS attacks, requiring both sensitive perception of network changes when DDoS occurs and reducing monitoring overhead to adapt to network resource constraints. In this paper, we propose a novel DDoS incident monitoring mechanism that uses macro-level network traffic behavior as a monitoring anchor to detect subtle malicious behavior indicative of the existence of DDoS traffic in the network. This behavior feature can be abstracted from our designed traffic matrix sample by aggregating continuous IPv6 traffic. Compared to IPv4, the fixed-length header of IPv6 allows more efficient packet parsing in preprocessing. As the decision core of monitoring, we construct a lightweight Binary Convolution DDoS Monitoring (BCDM) model, compressed by binarized convolutional filters and hierarchical pooling strategies, which can detect the malicious behavior abstracted from input traffic matrix if DDoS traffic is involved, thereby signaling an ongoing DDoS attack. Experiment on IPv6 replayed CIC-DDoS2019 shows that BCDM, being lightweight in terms of parameter quantity and computational complexity, achieves monitoring accuracies of 90.9%, 96.4%, and 100% when DDoS incident intensities are as low as 6%, 10%, and 15%, respectively, significantly outperforming comparison methods.

We present Recurrent Fitting (ReFit), a neural network architecture for single-image, parametric 3D human reconstruction. ReFit learns a feedback-update loop that mirrors the strategy of solving an inverse problem through optimization. At each iterative step, it reprojects keypoints from the human model to feature maps to query feedback, and uses a recurrent-based updater to adjust the model to fit the image better. Because ReFit encodes strong knowledge of the inverse problem, it is faster to train than previous regression models. At the same time, ReFit improves state-of-the-art performance on standard benchmarks. Moreover, ReFit applies to other optimization settings, such as multi-view fitting and single-view shape fitting. Project website: https://yufu-wang.github.io/refit_humans/

Abstract Best‐effort service model of traditional routing is gradually hard to meet the personalized demands under the rapid development of network technologies (e.g. 5G and IPv6). Therefore, service customization should be considered. In this work, a service customized routing mechanism based on deep learning in IPv6 network is proposed, which includes deep learning‐based service customization module, reliability evaluation module, and routing calculation module. The first module uses neural network to learn the complex service customization function, which can quickly output win‐win customized service strategies based on user demands. The second module can quantify the reliability of service routing paths, where not only the link status of IPv6 Neighbor Unreachable Detection (NUD) is considered, but also propose link performance weights to ensure the reliability of differentiated service performance. The third module uses the gray wolf optimization algorithm to calculate an optimal routing path to forward services with the customized strategies as the constraints and the maximum reliability and minimum cost as the goal. Finally, the mechanism is tested on the IPv6 Source Address Validation Improvement (SAVI) platform, which can reduce the execution time by 12.25% and improve the average routing reliability, user and ISP satisfaction by 9.0%, 40.45% and 7.4%, respectively.

Animals are diverse in shape, but building a deformable shape model for a new species is not always possible due to the lack of 3D data. We present a method to capture new species using an articulated template and images of that species. In this work, we focus mainly on birds. Although birds represent almost twice the number of species as mammals, no accurate shape model is available. To capture a novel species, we first fit the articulated template to each training sample. By disentangling pose and shape, we learn a shape space that captures variation both among species and within each species from image evidence. We learn models of multiple species from the CUB dataset, and contribute new species-specific and multi-species shape models that are useful for downstream reconstruction tasks. Using a low-dimensional embedding, we show that our learned 3D shape space better reflects the phylogenetic relationships among birds than learned perceptual features.

Software Defined Networking (SDN) decouples the control plane from the data plane, which increases network scalability and flexibility. But malicious applications on SDN controller can cause the entire network to crash. So, we design a reputation and incentive mechanism on SDN to reduce application's malicious access. In the proposed module, first of all, the application behavior is analyzed and the malicious accesses are identified, which are used to build the reputation and incentive mechanism. Second, the analysis results of the application behavior are combined through beta probability density to obtain the reputation rating. The reward or punishment will be given based on the behavior and reputation of the application under the selected social strategy. Simulation results show that the system can accurately identify malicious behavior and reduce malicious requests, with an acceptable runtime overhead about 300 microseconds.