Non-destructive detection methods based on molecular vibrational spectroscopy are pivotal in fields such as analytical chemistry and medical diagnostics. Recent advances have integrated deep learning with vibrational spectroscopy, significantly enhancing spectral recognition accuracy. However, these methods often rely on large annotated spectral datasets, limiting their general applicability. To address this limitation, we propose a novel approach, Global and Local Semantics Mining (GLSM), which leverages self-supervised learning to capture the global and local semantic information of unlabeled spectra, obviating the need for extensive annotated data. We devise two proxy tasks: global semantic mining and local semantic mining. The global semantic mining task is based on the premise that different views of the same spectrum can be mutually transformed, enabling the model to capture domain-invariant features across various perspectives and thereby develop a global understanding of the spectral data. This, in turn, enhances the model's robustness to variations in peak positions. Meanwhile, the local semantic mining task posits that noisy spectra can be reconstructed into noise-free spectra, thereby facilitating the extraction of local patterns and fine-grained details, such as subtle variations in peak intensities. By combining both self-supervised tasks, our model effectively captures the global and local semantic information of the spectrum. The pre-trained model can be fine-tuned with a limited amount of labeled homologous or heterologous spectral data for semi-supervised or transfer learning-based spectral classification. Extensive experiments on three datasets in semi-supervised and transfer learning-based spectral recognition tasks comprehensively validate the effectiveness of our GLSM method, demonstrating its significant potential for real-world spectral analysis applications.

Recent advances in the industrial inspection of textured surfaces—in the form of visual inspection—have made such inspections possible for efficient, flexible manufacturing systems. However, establishing a unified manual-feature-based inspection model for homogeneous and nonregularly textured surfaces presents an enormous challenge. Furthermore, in real industrial scenarios, collecting and labeling sufficient defective samples is impracticable due to the scarcity of defects and the endless variety of defect types, thus limiting the performance of supervised deep learning methods. To address these challenges, we propose an unsupervised feature memory rearrangement network (FMR-Net) to accurately detect various textural defects simultaneously. Consistent with mainstream methods, we adopt the idea of background reconstruction; however, we innovatively utilize artificial synthetic defects to enable the model to recognize anomalies, while traditional wisdom relies only on defect-free samples. First, we employ an encoding module to obtain multiscale features of the textured surface. Subsequently, a contrastive-learning-based memory feature module (CMFM) is proposed to obtain discriminative representations and construct a normal feature memory bank in the latent space, which can be employed as a substitute for defects and fast anomaly scores at the patch level. Next, a novel global feature rearrangement module (GFRM) is proposed to further suppress the reconstruction of residual defects. Finally, a decoding module utilizes the restored features to reconstruct the normal texture background. In addition, to improve inspection performance, a two-phase training strategy is utilized for accurate defect restoration refinement, and we exploit a multimodal inspection method to achieve noise-robust defect localization. We verify our method through extensive experiments and test its practical deployment in collaborative edge–cloud intelligent manufacturing scenarios by means of a multilevel detection method, demonstrating that FMR-Net exhibits state-of-the-art inspection accuracy and shows great potential for use in edge-computing-enabled smart industries. Note to Practitioners—Most conventional visual inspection methods rely on supervised training and consequently require a large amount of labeled data and can detect only specific types of texture defects. In contrast, the proposed FMR-Net is a robust model for the simultaneous and accurate inspection of textured surfaces for various defects that does not require any real labeled defect samples. Furthermore, this model can also support a different fine-grained detection method that is very suitable in the edge computing paradigm. These two characteristics are both extremely important for practical industrial applications. To the best of our knowledge, this is the first unsupervised edge intelligent vision inspection framework. As such, it can provide inspiration and serve as a reference for intelligent industry.

Hybrid active contour models with the combination of region and edge information have attracted great interests in image segmentation. To the best of our knowledge, however, the theoretical foundation of these hybrid models with level set evolution is insufficient and limited. More specifically, the weighting factors of their energy terms are difficult to select and are often empirically determined without definite theoretical basis. This problem is particularly prominent in the case of multi-object segmentation when more level set functions must be computed simultaneously. To cope with these challenges, this paper proposes a new level set approach for constructing hybrid active contour models with reliable energy weights, where the weights of region and edge terms can be constrained by the optimization condition deduced from the proposed method. It can be regarded as a general approach since many existing region-based models can be easily used to construct new hybrid models using their equivalent two-phase formulations. Some representative as well as state-of-the-art models are taken as examples to demonstrate the generality of our method. The respective comparative studies validate that under the guidance of the optimization condition, segmentation accuracy, robustness, and computational efficiency can be improved compared with the original models which are used to construct the new hybrid ones.

The active contour models with level set evolution have been visited with a vast number of methods for image segmentation. They can be mainly classified into region-based and edge-based models, and it has been validated that the hybrid variants combining both region and edge information can improve the segmentation performance. However, to the best of our knowledge, the theoretical foundation of collaboration mechanism between the region and the edge information is limited. Specifically, most existing hybrid models are just combining all the energy terms together, resulting in great challenges of choosing an appropriate weight coefficient for each term and accommodating different modalities of imaging. To overcome these difficulties, this paper proposes a region and edge synergetic level set framework named RESLS. It provides an approach to construct new hybrid level set models using a normalized intensity indicator function that allows the region information easily embedding into the edge-based model. In this case, the energy weights of region and edge terms can be constrained by the global optimization condition deduced from the framework. Some representative as well as state-of-the-art models are taken as examples to demonstrate the generality of our method. The experiments validate that under the guidance of the optimization condition, the weighting parameter of each term can be reliably chosen. Meanwhile, the segmentation accuracy, robustness, and computational efficiency of RESLS can be improved compared with its component models.

Online multi-object tracking remains a difficult problem in complex scenes because of inaccurate detections, frequent occlusions by clutter or other objects, similar appearances of different objects, and other factors. In this paper, we propose a hybrid data association strategy combined with instance segmentation and online feature learning to handle these difficulties effectively. First, we utilize an instance segmentation algorithm to separate each object from other objects and backgrounds in a pixel-to-pixel manner, which will help resolve typical difficulties in multi-object tracking, such as ID-switches and track drifting. Moreover, we propose a local-to-global hybrid data association strategy to take advantage of the superiorities of both online and batch data association methods. The local data association between observations in consecutive frames reduces the computational complexity, hence ensuring the efficiency of online tracking. The global data association complements the local data association by integrating multiple video frames, thus alleviating the fragmented tracklets. Last, to improve the appearance discriminability and make it more robust in dealing with appearance variations during tracking, a lightweight semantic-preserving hashing algorithm has been proposed to learn compact hash codes online. Experiments with the MOT17 Challenge dataset demonstrate the superior performance of the proposed approach over other state-of-the-art batch and online tracking methods.

Non-cooperative 3D face reconstruction is very significant in the area of intelligent security. According to non-cooperative 3D face reconstruction, the non-complete information fusion of multi-view face images can be realized to get a more complete face. This paper proposes a non-cooperative 3D face reconstruction method. A multimedia sensor network is employed to detect a person and get face images from different views. View-based active appearance models (View-based AAM) then helps to extract feature points and estimate probable pose angle. A new self-adaptive 3D morphable model based multi-view face geometry reconstruction method is designed to generate a 3D face model with particle swarm optimization (PSO). As the initial pose estimation is not accurate, particle swarm optimization is also used to regulate pose estimation results for optimizing 3D reconstruction result. “Mirror” strategy is employed to difine the invisible part of the face based on the mirror image of the visible part for texture mapping. Experiments have shown that the proposed method can achieve the non-cooperative 3D reconstruction efficaciously.

Sensor nodes deployment is very crucial for wireless sensor networks (WSNs). Current methods are apt to enlarge the coverage by achieving a nearly even deployment with similar density in the whole network. However, in some specific applications, the even distribution may not satisfy the sensing requirements. This paper proposes a virtual force directed coevolutionary particle swarm optimization (VFCPSO) algorithm, which uses a combined objective function to achieve the tradeoff of coverage and energy consumption. By considering deployment as an optimization problem, VFCPSO is more reliable and flexible for WSNs, since it can satisfy the combined requirements instead of only enlarging coverage. For investigating the performance of different paradigms, centralized VFCPSO is extended to distributed VFCPSO, heterogeneous hierarchical VFCPSO and homogeneous hierarchical VFCPSO (Homo-H-VFCPSO), and the solution of preferential deployment in interested region is also analyzed. Simulation results show that the Homo-H-VFCPSO has the best performance, i.e., it is more efficient than other three VFCPSO algorithms and the VF-style algorithms in terms of computation time, coverage and efficient moving energy consumption. It is obvious that the Homo-H-VFCPSO has good global searching ability and scalability, and it can rapidly and effectively achieve the sensor nodes deployment in WSNs.

Energy constraint is an important issue in wireless sensor networks. This paper proposes a distributed energy optimization method for target tracking applications. Sensor nodes are clustered by maximum entropy clustering. Then, the sensing field is divided for parallel sensor deployment optimization. For each cluster, the coverage and energy metrics are calculated by grid exclusion algorithm and Dijkstra's algorithm, respectively. Cluster heads perform parallel particle swarm optimization to maximize the coverage metric and minimize the energy metric. Particle filter is improved by combining the radial basis function network, which constructs the process model. Thus, the target position is predicted by the improved particle filter. Dynamic awakening and optimal sensing scheme are then discussed in dynamic energy management mechanism. A group of sensor nodes which are located in the vicinity of the target will be awakened up and have the opportunity to report their data. The selection of sensor node is optimized considering sensing accuracy and energy consumption. Experimental results verify that energy efficiency of wireless sensor network is enhanced by parallel particle swarm optimization, dynamic awakening approach, and sensor node selection.

Target tracking is an important task addressed in wireless sensor network (WSN), because the resources of WSN are limited. A distributed lightweight particle filter algorithm is proposed to achieve robust target tracking with low resource requirement in the WSN consisting of acoustic sensor nodes. In the proposed algorithm, each sensor node carries out partial particle filter with its local data and the neighbor nodes' data in distributed manner. Then local results of selected sensor nodes are fused to make final decision. To simplify the computation of particle filter, lightweight sampling and resampling schemes are introduced, where a simple range-free algorithm is adopted to restrict potential area of target's location. The experimental results verify that the proposed distributed lightweight particle filter algorithm can effectively achieve target tracking in WSN with low resource consumption. Compared to previous target localization algorithms, such as maximum likelihood estimation and centralized particle filter, the proposed algorithm has outstanding performance in accurate target tracking and low resources consumption.

A wireless sensor network (WSN) is a powerful unattended distributed measurement system, which is widely used in target surveillance because of its outstanding performance in distributed sensing and signal processing. This paper introduces a multiview visual-target-surveillance system in WSN, which can autonomously implement target classification and tracking with collaborative online learning and localization. The proposed system is a hybrid system of single-node and multinode fusion. It is constructed on a peer-to-peer (P2P)-based computing paradigm and consists of some simple but feasible methods for target detection and feature extraction. Importantly, a support-vector-machine-based semisupervised learning method is used to achieve online classifier learning with only unlabeled samples. To reduce the energy consumption and increase the accuracy, a novel progressive data-fusion paradigm is proposed for online learning and localization, where a feasible routing method is adopted to implement information transmission with the tradeoff between performance and cost. Experiment results verify that the proposed surveillance system is an effective, energy-efficient, and robust system for real-world application. Furthermore, the P2P-based progressive data-fusion paradigm can improve the energy efficiency and robustness of target surveillance.

Rotating machinery anomaly detection is of paramount significance for industries to prevent catastrophic breakdown and improve productivity and personnel safety. The kernel classifier support vector machine (SVM) has shown excellent performance towards this purpose, but it is difficult to optimize relevant hyper-parameters. In this paper, we propose a new anomaly detection approach by merging Gaussian process classifiers (GPCs) and bootstrap methods. GPCs are Bayesian probabilistic kernel classifiers and provide a well established Bayesian framework to determine the optimal or near optimal kernel hyper-parameters. They are largely unexplored for anomaly detection applications; consequently we take the initiatives to investigate GPCspsila performance in these scenarios. Bootstrap methods are incorporated to improve GPCspsila performance for small machinery anomaly samples by resampling at random. The proposed approach is evaluated on a motor testbed and wavelet packet is utilized to perform vibration analysis. Experiment results show bootstrap GPCs are highly effective and outperform GPCs and SVM with cross validation for anomaly detection. Moreover GPCs also prove to outperform SVM. Thus the proposed approach is promising for rotating machinery anomaly detection.

To implement visual target classification, this paper proposes a collaborative statistical learning algorithm for online support vector machine(SVM) classifier learning in wireless multimedia sensor network (WMSN). For achieving robust target classification, classifier learning should be carried out iteratively for updating classifiers according to various situations. Because only unlabeled samples can be acquired, semi-supervised learning is desired to make full use of unlabeled samples. According to the restrict limitation in energy and bandwidth, the proposed algorithm incrementally implement classifier learning with the selected features from multiple sensor nodes, where rough set based feature reduction is used for retaining most of the intrinsic information. Furthermore, some metrics are introduced to evaluate the effectiveness of the samples in specific sensor nodes, and a sensor node selection strategy is also proposed to reduce the impact of inevitable missing detection and false detection. Experimental results demonstrate that the collaborative statistical learning algorithm can effectively implement target classification in WMSN. With the rough set based feature reduction, the proposed algorithm has outstanding performance in energy efficiency and time cost.