In this work, we aim to reconstruct the 3D shape of an indoor scene from a single view, which includes multiple objects and the background. This task is challenging for existing methods since those instances of indoor scenes regularly occlude each other and contain diverse topologies. To address this, we propose a novel framework, ISDNet, to adaptively separate mixed instances and perform topology-aware reconstruction. Specifically, Specifically, ISDNet consists of two cascaded subnetworks: an instance separation module (ISM) and an instance deformation module (IDM). The ISM learns to separate occluded objects through stepwise sampling, inferring clean features for each instance. On the basis of these features, IDM generates an instance-topology-aware template and deforms it with learned offsets to reconstruct detailed geometry. Quantitative and qualitative experiments on the SUNRGB-D and 3D-FRONT datasets demonstrate that ISDNet outperforms the state-of-theart methods in terms of local details and overall shapes.

Existing point cloud completion methods rely on extracting latent codes from a partial point cloud to reconstruct a complete structure. However, the complexity of the partial point clouds, making the completion results of such methods less satisfactory, especially in long-distance (away from partial point cloud) areas. To tackle this challenge, we propose a point cloud completion network via heuristic structure growing (HSG-Net), which progressively completes the close-distance structure through an iterative heuristic structure growth strategy. Particularly, a novel data preprocessing (DP) method is proposed to obtain ground truth (GT) with specific structural integrity, guiding the network to learn close-distance structural information. In addition, the proposed consistency constraint displacement module (CCDM) is employed to fulfill structure growth, and a feature memory module (FMM) further enhances the quality of the grown structure. Furthermore, a proposed local information generator is used to further refine the structure-grown point cloud, fetching the final result. Extensive quantitative and qualitative results demonstrate that our HSG-Net outperforms the state-of-the-art methods.

Massive multiple-input–multiple-output (MIMO) systems provide efficient connectivity services for a large number of industrial Internet of Things (IoT) devices. To support numerous IoT devices, this article considers downlink hybrid architecture massive MIMO systems with phase shifters-based and switches-based cases, in which the data stream is first sent to the digital domain and then processed by the analog domain. zero-forcing (ZF) receivers are considered under the assumption of perfect channel state information (CSI). For hybrid architecture MIMO systems under quantized phase shifters and switches, we derive asymptotic approximations of the spectrum efficiency (SE) for massive MIMO systems using random matrix theory and then we propose an iterative algorithm to generate corresponding analog processing matrix, which can be used arbitrary quantized phase shifters. With the help of derived theoretical SE and realistic power consumption model, we further evaluate the total energy efficiency (EE) and provide insights into the tradeoffs between the SE and total EE. Finally, Monte Carlo simulation results are presented to substantiate our theoretical analysis results and showcase that the hybrid architecture provides higher total EE compared to existing all-digital architectures, but its achievable SE is lower than that of the hybrid architecture. Furthermore, the results also show that the switchbased hybrid architecture have ability to achieve the enhanced EE performance with large number of base station antennas and low-signal-to-noise ratio regimes compared to quantized phase shifters.

Point clouds obtained from 3D scanning are typically incomplete, noisy, and sparse. Previous completion methods aim to generate complete point clouds, while taking into account the densification of point clouds, filling small holes, and proximity-to-surface, all through a single network. After revisiting the task, we propose SDNet, which disentangles the task based on the spatial characteristics of point clouds and formulates two sub-networks, a Dense Refiner and a Missing Generator. Given a partial input, the Dense Refiner produces a dense and clean point cloud, as a more reliable partial surface, which assists the Missing Generator to better infer the remaining point cloud structure. To promote the alignment and interaction across these two modules, we propose a Cross Fusion Unit with designed Non-Symmetrical Cross Transformers to capture geometric relationships between partial and missing regions, contributing to a complete, dense and well-aligned output. Extensive quantitative and qualitative results demonstrate that our method outperforms the state-of-the-art methods.

Model migration can accelerate model convergence during federated learning on the Internet of Things (IoT) devices and reduce training costs by transferring feature extractors from fast to slow devices, which, in turn, enables sustainable computing. However, malicious or lazy devices may migrate the fake models or resist sharing models for their benefit, reducing the desired efficiency and reliability of a federated learning system. To this end, this work presents a blockchain-based model migration approach for resource-constrained IoT systems. The proposed approach aims to achieve secure model migration and speed up model training while minimizing computation cost. We first develop an incentive mechanism considering the economic benefits of fast devices, which breaks the Nash equilibrium established by lazy devices and encourages capable devices to train and share models. Second, we design a clustering-based algorithm for identifying malicious devices and preventing them from defrauding incentives. Third, we use blockchain to ensure trustworthiness in model migration and incentive processes. Blockchain records the interaction between the central server and IoT devices and runs the incentive algorithm without exposing the devices’ private data. Theoretical analysis and experimental results show that the proposed approach can accelerate federated learning rates, reduce model training computation costs to increase sustainability, and resist malicious attacks.