Recent advances in 3D Gaussian Splatting have shown remarkable potential for novel view synthesis. However, most existing large-scale scene reconstruction methods rely on the divide-and-conquer paradigm, which often leads to the loss of global scene information and requires complex parameter tuning due to scene partitioning and local optimization. To address these limitations, we propose MixGS, a novel holistic optimization framework for large-scale 3D scene reconstruction. MixGS models the entire scene holistically by integrating camera pose and Gaussian attributes into a view-aware representation, which is decoded into fine-detailed Gaussians. Furthermore, a novel mixing operation combines decoded and original Gaussians to jointly preserve global coherence and local fidelity. Extensive experiments on large-scale scenes demonstrate that MixGS achieves state-of-the-art rendering quality and competitive speed, while significantly reducing computational requirements, enabling large-scale scene reconstruction training on a single 24GB VRAM GPU.

Despite the great progress of deep learning and generative adversarial networks, face frontalization (i.e., profile-to-frontal synthesis) and profile (i.e., non-frontal) face recognition still remain challenging tasks under uncontrolled environments. In this study, we propose three efficient practices to improve the performance of profile-to-frontal face synthesis and recognition. Firstly, the identity preserving module is embedded to constrain synthesized frontal images similar to true frontal faces in feature space. Secondly, facial consistency loss is employed to reduce the artifact of the generated frontal face in pixel space. Lastly, the multi-model embedded scheme enhances the representation learning through diverse facial features extracted by multiple facial feature extractors. The proposed practices are general, though specifically deployed to CR-GAN in this study for performance verification. Experimental results on Multi-PIE and VGGFace2 demonstrate that the proposed practices qualitatively generate more realistic photography frontal faces and quantitatively obtain better face recognition accuracy.

In this paper, we study$H_{\infty} $pinning synchronization control problem for Markovian intermittent coupled neural networks with time-varying coupling strengths, and the influence of multiplicative noises on actuators is also considered. A set of random variables of known mathematic expectations and variances is introduced to describe stochastically changing interactions among neural networks, and a two-states Markov chain is used to characterize the intermittently time-varying interactive relationship between neural network and its neighboring neural networks. By analyzing stability and$H_{\infty} $disturbance attenuation performance of the obtained synchronization error systems, two sufficient conditions are derived to ensure global synchronization and$H_{\infty} $global synchronization of the controlled networks respectively. Then, the pinning synchronization controllers are designed, and we can change control cost by varying the number of pinned nodes. Finally, an algorithm resulting the set of optimal pinned nodes which minimizes the$H_{\infty} $disturbance attenuation performance index is provided, and a numerical example illustrating the effectiveness of the obtained theoretical results is given.

Recent advances in deep neural networks, especially in generative adversarial networks (GAN), have shown remarkable progress in face image generations. However, most of the existing face image generators can only synthesize random face images, but are not able to control the attributes of the generated face images. Though conditional GAN based methods can manipulate the attributes to some extent, but can only generate low-resolution face images up to 256 × 256. In this study, based on StyleGAN, one of the state-of-the-art image generators for synthesizing high-quality face images, we propose a simple but efficient approach to generate high-resolution and hyper-realistic face images with any desired attribute. By training an attribute classifier to assign attribute labels to given synthesized face images, we build the links between latent vectors and face attributes. In such a way, the latent vectors can be grouped into different clusters, one cluster corresponding to one face attribute, respectively. We then extract the prototypes for the clusters, which are used to control the attribute of the generated face image. Extensive experiments demonstrate the effectiveness of the proposed approach for high-quality face image generation with predefined attributes.

In this article, the resilient anti-disturbance control is studied for uncertain turbofan system subject to unmeasured states and multiple disturbances. Based on four kinds of disturbances in the addressed system, some disturbances are described as an external system by using available information while the others are assumed to be energy-bounded, which are included in the system dynamics, output measurement, and controlled output, simultaneously. Initially, a state observer and a disturbance observer are jointly constructed to estimate the unmeasured state and unknown disturbance. The estimation on disturbance is used in the feedforward controller to reject the disturbances and the state estimation is applied to the resilient output feedback controller, which guarantee that the closed-loop system is asymptotically stable with the L2-L∞performance, and enhance the robustness of the uncertain turbofan system. Then, the Lyapunov stability theory and linear matrix inequality (LMI) technology are combined to obtain the algorithms on checking the controller gain and observer one. Finally, the effectiveness of our proposed methods is shown by using some numerical simulations.

Source location privacy, one of the core contents of Wireless Sensor Network (WSN) security, has a significant impact on extensive application of WSNs. In this paper, a novel location privacy protection routing scheme called Energy Balanced Branch Tree (EBBT) is proposed by using multibranch and fake sources. This scheme has three phases. In the first place, the data of the source are randomly sent to a certain intermediate node. Then, a minimum hop routing (MHR) from the intermediate node to the base station is formed. Then, branch paths with fake sources are generated dynamically from some nodes on the MHR path. Finally, a tree‐shaped structure from real source nodes and fake source nodes to the base station is achieved. In difference to the previous schemes, the location of the real source in the EBBT scheme does not affect the location and the number of fake sources. During the formation of the tree‐shaped multibranch paths, the residual energy of nodes is considered sufficiently, and the control of the direction of each branch path is also involved. The influence of the number and length of branches on the network lifetime and network security is also investigated. Experimental results show that the proposed algorithm has the advantages of long network security period and lifetime, as well as high path diversity. Our simulation further illustrates that the EBBT scheme has favorable privacy of the source location without changing the network lifetime.

Visible and near-infrared (VIS-NIR) face recognition remains a challenging task due to distinctions between spectral components of two modalities. Inspired by the CycleGAN, this paper presents a method aiming to translate between VIS and NIR face images. To achieve this, we propose a new facial feature embedded CycleGAN. Firstly, to learn the particular feature while preserving common facial representation between VIS and NIR domains, we employ a general facial feature extractor (FFE) to extract effective features. Herein the MobileFaceNet is pre-trained on a VIS face database and serves as the FFE. Secondly, the domain-invariant feature learning is enhanced by proposing a new pixel consistency loss. Lastly, we establish a new WHU VIS-NIR database including varies in face rotation and expressions to enrich the training data. Experimental results on the Oulu-CASIA and our WHU VIS-NIR databases show that the proposed FFE-based CycleGAN (FFE-CycleGAN) outperforms some state-of-the-art methods and achieves 96.5% accuracy.