Existing 3D human pose estimation (HPE) methods typically prioritize lifting 2D pose coordinates to 3D but tend to underemphasize the importance of generalizing under real-world conditions with noisy 2D inputs from off-the-shelf 2D detectors. In this paper, we introduce Graph Attention Fusion Network (GAtFuN), a novel motion-aware framework that integrates our spatial and temporal graph attention mechanisms to explicitly model joint velocities and motion transformations, resulting in more stable and coherent 3D pose predictions despite being trained with the same dataset pipeline as other SOTA methods. GAtFuN achieves a 7.8% improvement in MPJPE over the current SOTA on the Human3.6M dataset and a 1.9% improvement on the MPI-INF-3DHP dataset, while demonstrating more robust performance on the 3DPW dataset in the wild.

Existing few-shot cross-component fault diagnosis methods primarily focus on the correlation between input data and fault classes, neglecting causal relationships. This limits the model's ability to separate and eliminate confounding factors, limiting the improvement of cross-component prediction accuracy. To address this issue, this paper proposes a Causal Intervention Network for Cross-Component Few-Shot Fault Diagnosis (CINet), which constructs a causal structure model to perform causal decomposition, extracting and decoupling the instrumental variable, confounding variable, and adjustment variable from vibration signals, thus enabling the modeling of cross-component causal relationships, which enhances diagnostic accuracy under few-shot conditions. Specifically, the CINet is composed of three main modules: a feature encoding module, a causal disentanglement module, and a relation metric module, jointly optimizing the fault diagnosis and relation metric selection loss functions through multi-task learning. Experimental results on multiple fault diagnosis datasets demonstrate that the CINet significantly outperforms existing methods, especially in handling cross-component fault diagnosis problems, particularly in few-shot scenarios, by better capturing causal relationships and improving prediction accuracy and model interpretability.

Aiming at the problem of task offloading in multiaccess edge computing (MEC) scenarios, this paper proposes a meta-reinforcement learning (Meta-RL)-based computational task offloading method. The algorithm adopts a two-layer architecture: the inner layer models the task offloading process as a Markov Decision Process (MDP), designs a reward function based on task latency and energy consumption, and designs a task offloading algorithm based on Proximal Policy Optimization (PPO) to make offloading decisions for each task in a single scenario and optimize the offloading performance within the scenario. The outer layer introduces the Meta-RL mechanism to optimize the initial parameters of the inner-layer neural network and learns multiple MDPs based on gradient descent to generate neural network parameters that can be applied to the intelligence of each scenario, so that the proposed algorithm can adapt to the offloading scenarios quickly. The proposed algorithm can quickly adapt to each offloading scenario. Simulation results show that the proposed algorithm improves the average cost by 14.7 % and 20.51 % compared with PPO and DDPG.

Graph Convolutional Networks (GCNs) have proven to be highly effective for skeleton-based action recognition, primarily due to their ability to leverage graph topology for feature aggregation, a key factor in extracting meaningful representations. However, despite their success, GCNs often struggle to effectively distinguish between ambiguous actions, revealing limitations in the representation of learned topological and spatial features. To address this challenge, we propose a novel approach, Gaussian Topology Refinement Gated Graph Convolution (G3CN), to address the challenge of distinguishing ambiguous actions in skeleton-based action recognition. G3CN incorporates a Gaussian filter to refine the skeleton topology graph, improving the representation of ambiguous actions. Additionally, Gated Recurrent Units (GRUs) are integrated into the GCN framework to enhance information propagation between skeleton points. Our method shows strong generalization across various GCN backbones. Extensive experiments on NTU RGB+D, NTU RGB+D 120, and NW-UCLA benchmarks demonstrate that G3CN effectively improves action recognition, particularly for ambiguous samples.

Panoramic driving systems are crucial for autonomous driving but face challenges in real-time performance and reliability. This paper proposes an end-to-end, multi-tasking MtpNet that reduces latency and enhances detection accuracy. The convolution was upgraded using the Efficient Layer Aggregation Network, and precise multi-task loss functions and more effective training strategies were devised. Our results demonstrate improved performance in small object detection, partial occlusion handling, and drivable area segmentation. The recall of the traffic object detection is 1.3% higher than that of the state-of-the-art model, reaching 94.1%, the mAP50is 6.4% higher, reaching 89.8%, and the mIoU of the drivable area segmentation is 2.7% higher, reaching 95.9%. Additionally, the accuracy of lane detection reached 88.7%. The visual comparison using three datasets TuSimple, CityScapes, and CULane demonstrates that MtpNet has good detection segmentation and strong robustness under various conditions. Codes are available at https://github.com/ErLinErYi/mtpnet

The decoder is a key component in deep networks for the semantic segmentation of street views. The existing methods rely on the limited receptive field for feature extraction without considering the contextual information, which could lead to errors in understanding complex scenes. Moreover, a balance of contextual information and computational cost must be considered to meet the needs of real-world applications. To address these problems, we introduce a Spatial Plaid Attention Decoder network, which uses a lightweight decoder with Spatial Plaid Attention to perform highly efficient operations for semantic segmentation. With approximately 4 million parameters (9.75% of the UPerNet), our decoder achieves state-of-the-art performance on public datasets such as Cityscapes and ADE20K, with 84.84% and 54.0% mIoU, respectively. In addition, our method reduces the total Flops by 34.95% and 32.85%, respectively. We demonstrate how contextual information helps the network in object recognition and how object features and contextual features contribute to the scene segmentation and recognition.

Unmanned aerial vehicles (UAV), due to their flexibility and extensive coverage, have gradually become essential for substation inspections. Wireless mesh networks (WMN) provide a scalable and resilient network environment for UAVs, where each node can serve as either an access point or a relay point, thereby enhancing the network’s fault tolerance and overall resilience. However, the UAV-WMN combined system is complex and dynamic, facing the challenge of dynamically adjusting node transmission power to minimize end-to-end (E2E) delay while ensuring channel utilization efficiency. Real-time topology changes, high-dimensional state spaces, and large solution spaces make it difficult for traditional algorithms to guarantee convergence and stability. Generic reinforcement learning (RL) methods also struggle with stable convergence. This paper introduces a new Lyapunov function-based proof to address these issues and provide a stable condition for dynamic control strategies. Then, we developed a specialized neural network power controller and combined it with the MATD3 algorithm, effectively enhancing the system’s convergence and E2E performance. Simulation experiments validate the effectiveness of this method and demonstrate its superior performance in complex scenarios compared to other algorithms.

Striking a balance between efficiency and accuracy is a challenge in the design and implementation of decoders. Accurate decoders often tend to be highly complex and computationally costly. This paper presents a novel decoder for semantic segmentation: Spatial Plaid Attention Decoder (SPADe). We propose a Spatial Plaid Attention module that performs an efficient local feature collection through spatial feature folding to help the model capture the local structure of objects while using long-range feature aggregation to consider the global structures efficiently and accurately. This makes SPADe a suitable choice for applications with limited resources. With a size of 7.3% UPerNet decoder, our SPADe obtains state-of-the-art performance with several popular backbones on public benchmarks. On Cityscapes and ADE20K, SPADe obtains 84.3% and $53.8 \% \mathrm{mIoU}$, while reducing the total GFlops by 32.8% and 70.5%, respectively. We also demonstrate that the effective design of SPADe allows it to capture long-range dependencies with a large receptive field. Implementation of the SPADe is available at github/SPADe.

As part of airport behavior decisions, the accurate prediction of airport delay is highly significant in optimizing flight takeoff and landing sequences. However, the combination of various influencing factors affects airport delay prediction strongly, which would bring severe challenges in prediction. This paper introduces the sequence-to-sequence network and proposes a multi-step regression prediction method for the airport delay (DA-BILSTM) to accurately predict the airport delay. Rather than only considering a single kind of airport delay influencing factors, we design an attention fusion network for learning the sequence and condition correlation features adaptively. Moreover, the Bayesian optimization algorithm is introduced to optimize DA-BILSTM’s hyperparameters. The method is applied individually to two datasets for predicting the airport’s delays. The experiment results show that the prediction performance of DA-BILSTM is better than many state-of-the-art methods including the autoregressive integrated moving average model (ARIMA), long short-term memory (LSTM), gated recurrent unit(GRU), CNN-BILSTM, and TS-LSTM. When using DA-BILSTM in the two datasets, the average MAE of airport delay prediction in the next 5 hours is about 10 minutes, and the average RMSE is 20 minutes.

All six papers in this special section engage in different streams but extremely relevant domain vectors of Cognitive Cyber-Physical Systems (CCPS) with artificial intelligence (AI) based solutions in medical informatics. Highlights recent trends in the scientific community and presents emergent technologies, implementations, applications concerning the CPSS. CPSS is witnessing rapid transformation as an interdisciplinary technology that blends physical components and computing devices to enable AI-based solutions. CCPS will be playing a significant role that integrates machine learning/AI techniques and resulting in dramatic improvements for medical informatics and the future of human-augmentation. CPHMS coordinates supervisory medical systems and medical resources everywhere; there is a great scope towards health consciousness and healthy society. Medical Cyber-Physical Systems (MCPS) in healthcare towards critical integration in network of medical devices. MCPS is the next generation computing that is comprised of tightly coupled computational and communication components of medical automation systems such as clinical decision, early detection of health infectious, disease prevention, rapid analysis of health hazards and so on. CCPS and MCPS research would be created new models, new design, and integration models for large scale systems in comprehensive, holistic medical automation systems. With recent enlargements in the big data processing, cognitive data science and AI, it is now possible to create even more realistic digital twins that properly model different operating situations and characteristics to process the medical intelligence systems.

Non-rigid point set registration is challenging when point sets have large deformations and different numbers of points. Examples of such point sets include human point sets representing complex human poses captured by different types of depth cameras. In this work, we present a probabilistic, non-rigid registration method to deal with these issues. Two regularization terms are used: key point correspondences and local neighborhood preservation. Our method detects key points in the point sets based on geodesic distance. Correspondences are established using a new cluster-based, region-aware feature descriptor. This feature descriptor encodes the association of a cluster to the left-right (symmetry) or upper-lower regions of the point sets. We use the Stochastic Neighbor Embedding (SNE) constraint to preserve the local neighborhood of the point set. Experimental results on challenging 3D human poses demonstrate that our method outperforms the state-of-the-art methods. Our method achieved highly competitive performance with a slight increase of error by 3.9% in comparison with the method using manually specified key point correspondences.

Due to the increasing popularity of 3D sensors, it has become easier and easier to obtain point cloud data. In many fields such as autonomous driving, how to fully extract the features of point cloud to better understand and perceive 3D scenes requires further research. Therefore, this paper proposes a novel end-to-end deep learning network for the features of disorder and irregularity of 3D point cloud data. Our network uses an encoder-decoder network architecture, and a three-layer structure is used in both stages. Each encoder layer consists of graph attention convolution and graph attention pooling. Graph attention convolution reflects the spatial distribution relationship in the neighborhood area. And graph attention pooling merges the spatial distribution information in the neighborhood into the feature of the sampling point. The experimental results show that our method achieves the best results in shape classification and has competitive performance in other tasks.

Urban regions are complicated functional systems that are closely associated with and reshaped by human activities. The propagation of online geographic information-sharing platforms and mobile devices equipped with the Global Positioning System (GPS) greatly proliferates proximate sensing images taken near or on the ground at a close distance to urban targets. Studies leveraging proximate sensing images have demonstrated great potential to address the need for local data in the urban land-use analysis. This paper reviews and summarizes the state-of-the-art methods and publicly available data sets from proximate sensing to support land-use analysis. We identify several research problems in the perspective of examples to support the training of models and means of integrating diverse data sets. Our discussions highlight the challenges, strategies, and opportunities faced by the existing methods using proximate sensing images in urban land-use studies.

Radiation therapy (RT) of lung tumors requires an accurate and real-time localization of the tumor while the patient is being treated. Tumor motion caused by breathing can impact dose delivery in patients with lung cancer, leading to poor disease management and damage to surrounding normal tissues. A major challenge in tumor tracking using fluoroscopic imaging is to track the tumor while it is occluded by the overlapping bones (ribs and spine). In this work, we propose a series of three modeling strategies combining template matching for initial probability estimation and Kalman filter for sequential measurement updates. The images were acquired using a fast-kV switching real-time fluoroscope utilizing a dynamic thorax motion phantom. We utilized three distinct physics-based state representations to predict the motion of the tumor: Velocity, Acceleration, and Spring Motion. The Kalman filter with Spring Motion physics performed best with the average RMSE rates of 0.44 mm improving template matching alone with RMSE 1.81 cm. Additionally, occlusion and noise were simulated with general improvements from the Spring Motion Kalman filter. The proposed method will enable more accurate and precise lung radiotherapy using existing hardware and workflow. Our future work is focused on the clinical implementation of this method.

The unrestricted open vocabulary and diverse substances of scenery images bring significant challenges to scene recognition. However, most deep learning architectures and attention methods are developed on general-purpose datasets and omit the characteristics of scene data. In this paper, we exploit the Attention Pyramid Module (APM) to tackle the predicament of scene recognition. Our method streamlines the multi-scale scene recognition pipeline, learns comprehensive scene features at various scales and locations, addresses the interdependency among scales, and further assists feature re-calibration as well as the aggregation process. APM is extremely light-weighted and can be plugged into existing network architectures in a parameter-efficient manner. By integrating APM into ResNet-50, we obtain a boost of top-1 accuracy by 3.54% on the benchmark dataset. Our comprehensive experiments demonstrate that APM achieves much improved performance comparing with the state-of-the-art attention methods using significantly less computation budget. Source code is provided on https://github.com/ZN-Qiao/APM.

Land-use (LU) classification of urban areas is conventionally achieved via field survey or remote sensing technologies, which is labor-intensive and time-consuming. With the wide development of social networks such as microblog and ubiquitous network access, images are captured by residents and tourists. In this letter, we propose a method for an automatic urban LU classification using geotagged images from public venues. Our method identifies the LU type depicted in those images that are extrapolated to the local regions bounded by street blocks. Experiments were conducted with geotagged photographs and Open Street Map of an urban area in London, U.K. It was demonstrated that the proposed method achieved overall 76.5% accuracy across five LU types. More importantly, our method demonstrated a greater performance in dealing with a mixture of LU types.

Urban function zoning is of great importance for urban structure optimization, urban resource allocation, and urban development planning. Since citizens usually act as a network of motion sensors of the city, their activities could reflect the environment around them. We considered taking advantage of VGI data to classify urban function zones. In this paper, we proposed a framework for automated urban function zoning which is based on VGI geo-tagged photos and OpenStreetMap (OSM) data. Through combining the high-level image features of geo-tagged photos with the road network data, we obtained the functional zoning map of the study area. The experiment result shows the effectiveness of the framework we proposed.

Polarimetric SAR classification is an effective approach in image understanding. This paper proposes a novel semantic method for classification of Polarimetric SAR data. The method combines superpixels and semantic model to benefit from both the object-oriented classification and the high-level semantic information. Firstly, pixels was grouped into superpixels via Simple Linear Iterative Clustering (SLIC). Secondly, the feature vector was generated within the superpixels by considering both polarimetric information and textures. To incorporate semantic information, the feature vectors were further processed via probabilistic Latent Semantic Analysis (pLSA). Finally, Supporting Vector Machine (SVM) was utilized to obtain classification results. The results were evaluated with respect to the accuracy of classification and spatial preservation. The results of this work were analyzed by means of RADARSAT-2 data.

Based on the traffic accident data of Beijing China in 2012, we combined with a variety of municipal administration data, used the geographically weighted regression (GWR) method to study spatial non-stationarity and heterogeneity of the traffic accidents, and analyzed the causes of space regional in traffic accident black spots. Experimental results demonstrated that: (1) The GWR model detects change of the coefficient and explains the significance of coefficient change as well. It is suitable for interpreting the distribution regularity and the causes of traffic accident black spots at the microscopic level. (2) By analyzing the importance of the factors on traffic accident black spots, we note the following order of factors according to the significance: weather, hospital, park, subway station, curvature of the roads and population density. The influence of each factor on traffic accident black spots differs in spatial heterogeneity characteristics.

The life of many wireless sensor networks is limited by their battery-based energy source. Recharging batteries from a distance by the wireless energy transferring technique could lift this restriction. However, how to deploy the mobile charging device requires further research. In this paper, we take the energy constraint of mobile wireless charger (MWC) into consideration and aim at minimizing the total energy consumption by it in recharging cycles. After formulating the optimization problem, we present the MMES-LME method based on the modified MAXMIN Ant System and equalization strategy with the constraint of MWC limited energy. The equalization strategy is presented to equalize the lifespan of all sensor nodes to avoid the untimely death of WSN and balance the consumption of MWC travelling energy and recharging energy. Our experimental results demonstrate improved performance in comparison to the greedy method and MM-LME method, which is based on MAX-MIN Ant System.

Digital signature schemes with additional properties have broad applications, such as in protecting the identity of signers allowing a signer to anonymously sign a message in a group of signers (also known as a ring). While these number-theoretic problems are still secure at the time of this research, the situation could change with advances in quantum computing. There is a pressing need to design PKC schemes that are secure against quantum attacks. In this paper, we propose a novel code-based threshold ring signature scheme with a leader-participant model. A leader is appointed, who chooses some shared parameters for other signers to participate in the signing process. This leader-participant model enhances the performance because every participant including the leader could execute the decoding algorithm (as a part of signing process) upon receiving the shared parameters from the leader. The time complexity of our scheme is close to Courtois et al.’s (2001) scheme. The latter is often used as a basis to construct other types of code-based signature schemes. Moreover, as a threshold ring signature scheme, our scheme is as efficient as the normal code-based ring signature.

Images captured in hazy or foggy weather conditions are seriously degraded by the scattering of atmospheric particles, which directly influences the performance of outdoor computer vision systems. In this paper, a fast algorithm for single image dehazing is proposed based on linear transformation by assuming that a linear relationship exists in the minimum channel between the hazy image and the haze-free image. First, the principle of linear transformation is analyzed. Accordingly, the method of estimating a medium transmission map is detailed and the weakening strategies are introduced to solve the problem of the brightest areas of distortion. To accurately estimate the atmospheric light, an additional channel method is proposed based on quad-tree subdivision. In this method, average grays and gradients in the region are employed as assessment criteria. Finally, the haze-free image is obtained using the atmospheric scattering model. Numerous experimental results show that this algorithm can clearly and naturally recover the image, especially at the edges of sudden changes in the depth of field. It can, thus, achieve a good effect for single image dehazing. Furthermore, the algorithmic time complexity is a linear function of the image size. This has obvious advantages in running time by guaranteeing a balance between the running speed and the processing effect.

Sparse representation has been applied to visual tracking. The visual tracking models based on sparse representation use a template set as dictionary atoms to reconstruct candidate samples without considering similarity among atoms. In this paper, we present a robust tracking method based on the inverse sparse group lasso model. Our method exploits both the group structure of similar candidate samples and the local structure between templates and samples. Unlike the conventional sparse representation, the templates are encoded by the candidate samples, and similar samples are selected to reconstruct the template at the group level, which facilitates inter-group sparsity. Every sample group achieves the intra-group sparsity so that the information between the related dictionary atoms is taken into account. Moreover, the local structure between templates and samples is considered to build the reconstruction model, which ensures that the computed coefficients similarity is consistent with the similarity between templates and samples. A gradient descent-based optimization method is employed and a sparse mapping table is obtained using the coefficient matrix and hash-distance weight matrix. Experiments were conducted with publicly available datasets and a comparison study was performed against 20 state-of-the-art methods. Both qualitative and quantitative results are reported. The proposed method demonstrated improved robustness and accuracy and exhibited comparable computational complexity.

Large-scale Internet applications provide service to end users with servers, which may be located at geographically distributed data centers. Users may require different delay constraints for different services. To meet the service delay requirements to end users, the data centers must provide enough server resources which incur a large amount of electricity and dollars cost. In this paper, we tackle the problem of minimizing electricity cost under diverse delay requirements of different services for different users in a multi-electricity-market environment. We propose two algorithms to reduce the electricity cost, taking into account the location diversity and the time diversity of electricity price. Our simulation results demonstrated that the proposed algorithms were effective in terms of the reduction of the electricity cost while satisfying the diverse delay constraints.

Hazy images hinder image understanding in many applications such as autonomous vehicle. In this paper, we propose an efficient method to improve image quality of hazy images. Our method estimates the transmission function based on a linear model that allows efficient computation and employs quadtree to search for a region that best represents the scatter of airlight. Experiments were conducted using publicly available images. It is demonstrated that our proposed method achieved comparable results to the state-of-the-art ones. In the estimation of sunlight radiance, the quadtree that integrates local brightness and gradient as well as spatial constraint provide a robust means to identify region of sky. Most significantly, our proposed method greatly improved the efficiency. When dealing with moderate and large size image, the improvement could be more than thirty-fold.

With the development of small unmanned aerial systems (SUAS), the flight safety issues in low altitude has become a research focus. Clearance survey is used for locating obstacles in the airspace to ensure flight safety. In this paper, we employ LiDAR to generate digital terrain model (DTM) for achieving clearance survey. In our method, filters are designed to eliminate the abnormal data points and grids are used for recognizing obstacles. According to the principle of flight risk minimization, a maximum elevation based method is proposed to produce the digital surface model (DSM) and realize the DTM extraction. Combining the derived DTM and DSM, the low altitude space are divided into zones to ensure the flight safety of SUAS. The experimental results demonstrate the effectiveness of our proposed method. By varying the window size of the filters for processing LiDAR datasets, further improvements can be achieved.

Registering terrestrial LiDAR data points faces many challenges. Limited overlap between data sets, occlusions, and data density divergence greatly affect the registration performance. In this paper, we propose a method to identify the homologous points in different LiDAR acquisitions and join small but related primitives to maximize the correspondence between two data sets. Experiments were conducted with data sets collected using a Riegl VZ-400 terrestrial laser scanner. Our method greatly reduces the matching complexity from raw point measurements to planar primitives to clusters of spatially related structure features. The automatic registration method is able to achieve satisfactory results that put sets of LiDAR point clouds into correspondence with great accuracy.

Entropy, alpha, and anisotropy (H/α̅/A) of Cloude decomposition are effective in polarimetric SAR image understanding and geophysical information inversion. As an incoherent target decomposition, the inner sample covariance matrix estimation severely affects the estimated parameters. The contradiction between details preservation and accurate parameters estimation is still a challenge task. In this article, we propose adaptive coherency matrix estimation based on local heterogeneity coefficients, and utilize it to parameters estimation of Cloude decomposition. The results were evaluated with respect to details preservation and the accuracy of parameters estimation. The results of this work were analyzed by means of AIRSAR data.

Abstract Clustering methods have been developed to improve network life of wireless sensor network (WSN), yet the dynamic nature of sensor clusters and limited memory and processing power make security a much more challenging problem, and most conventional cryptography methods are ill suited to WSNs. In this paper, we propose a novel encryption method to secure data transmission in WSN with dynamic sensor clusters. Our method leverages elliptic curve cryptography algorithm to generate binary strings for each sensor and combines with node ID, distance to the cluster head, and the index of transmission round to form unique 176‐bit encryption keys. Using exclusive OR, substitution, and permutation operations, encryption and decryption are achieved efficiently. Compared with the state‐of‐the‐art methods, our simulation results demonstrated that the proposed method exhibited much improved network lifetime and reduced the energy consumption most evenly among all sensor nodes. More importantly, it overcame many security attacks including brute‐force attack, HELLO flood attack, selective forwarding attack, and compromised cluster head attack. Copyright © 2016 John Wiley & Sons, Ltd.

Polarimetric synthetic aperture radar (SAR) images usually contain a mixture of homogeneous and heterogeneous regions, which makes estimation of the coherency matrix a very challenging task. In this paper, we propose an adaptive coherency matrix estimation method that employs local heterogeneity coefficient and leverages the sample covariance matrix estimation to the homogeneous components and the fixed-point estimation to the heterogeneous components. Evaluations were conducted with synthetic polarimetric data and real-world SAR imagery, including UAVSAR, RADARSAT-2, and ESAR. Our experimental results demonstrated that the heterogeneity coefficient effectively characterizes the scattering property of ground objects, which enables adaptive estimation of the coherency matrix in high-resolution polarimetric SAR imagery. Our method was able to handle single- and multilook polarimetric SAR imagery gracefully. Compared with the sample covariance matrix estimator, the fixed-point estimator, and the Lee sigma filtering, our method achieved the best performance for retaining the spatial structure, suppressing speckles, and preserving polarimetric information of SAR imagery with different degrees of heterogeneity.

In this paper, we propose a multi-class boosting method (multiBoost.imb) to address difficulties of learning from imbalanced data set as well as employment of stable base learners. A random resampling strategy is incorporated to diversify the training data set and to recover balance among all classes. Extending AdaBoost by adding an error adjustment parameter, early termination in the training phase is avoided in multi-class scenarios. Experiments were conducted using three public face databases and two synthetic data sets. It is demonstrated that stable learners can be used in our ensemble method. In the multi-class problems, the ensemble overcomes the early termination even when stable learner is employed. It was evident that our method improves learning performance in all cases, especially when imbalance ratio is high. Comparison to the SMOTEboost and RUSboost also reveals the advantage of our method in handling multi-class, imbalanced face recognition problems.

Identifying hydrological features is important for urban planning and disaster assessment. Data spatial resolution poses challenges in automatic processing. In this letter, we present a novel spatially constrained model-driven clustering method that automatically detects and delineates water bodies in an urban area using airborne laser swath mapping (ALSM) data and imagery. Our method analyzes the modality of the sparseness histogram to decide the existence of water body, followed by clustering. Using the sparseness, clusters are decided by selecting candidate sites. In the iteration of clustering process, new sites are recruited within a close spatial vicinity of the boundary sites. Experiments were conducted using data sets from the city of New Orleans. Our method demonstrated superior robustness regardless of the density of ALSM sample and data discrepancy and very competitive accuracy in comparison with manual tracing, with an overall accuracy above 98%.

Tagline detection and indexing are challenging tasks due to complicated anatomical properties and imaging noise. In this paper, we will address the following two important issues in tagline detection: 1) an automatic method independent from imaging approaches with improved robustness and accuracy and 2) tagline indexing that matches taglines in task and reference images for postprocessing. Our method consists of two steps: First, a wavelet decomposition is performed on a tagged magnetic resonance (tMR) image. Subband correlation is used to dampen anatomical boundaries but enhance taglines. A tagline map is created by segmenting a reconstructed image using pseudowavelet reconstruction. Next, tagline pixels are grouped into clusters and isolated small line segments are eliminated. A snake method is then used to index and recover broken taglines. Our method has been validated with 320 tMR tongue images. Measurement of tagline accuracy was performed by computing tag pixel displacement. Without assumptions on tagline models, it detects taglines automatically. Comparison studies were conducted against the harmonic phase method. Our experiments resulted in a p-value of 1E-6 with one-way ANOVA, which indicates a significant improvement in accuracy and robustness.

We present a preprocessing method for automatic break line detection. Our method is given a set of edges (break lines and non-break lines) we eliminate the non-break line edges leaving only those with higher probability of being break lines. Our method is based on fusing IR images and LiDAR cloud points. In the first step, we apply the Canny edge detection algorithm to the IR images (producing a superset of break lines). Then we project the LiDAR points onto a 2-D plane, ignoring the set of points that are greater than a selected threshold (different elevation thresholds have been selected), which allows the footprints of some elevated structures to appear clearly in the set of projected points. Those structures that appear in both LiDAR points and IR images are used as references for registration of LiDAR cloud points with the IR images. After registration, we eliminate all the edges that appear in flat areas, which is achieved by applying a 3D filter to the LiDAR points.

To generate a DTM, measurements from above-ground features such as buildings, vehicles, and vegetation have to be classified and removed, which is nontrivial. The above-ground features present great challenges in conjunction with varying slopes of the ground. In this paper, we present a method to remove above-ground LiDAR measurements and generate DTMs by using adaptive window size according to the local gradients. Iterative construction measurements are performed until difference between two iterations are minimum. In our experiments, we apply our method to the LiDAR data acquired from the downtown region of New Orleans. It was demonstrated that the adaptive window method can remove most of the above-ground points effectively.

Fusion of images in the presence of noise is a challenging problem. Conventional fusion methods focus on aggregating prominent image features, which usually result in noise enhancement. To address this problem, we developed a wavelet-based, noise-aware fusion method that distinguishes signal and noise coefficients on-the-fly and fuses them with weighted averaging and majority voting respectively. Our method retains coefficients that reconstruct salient features, whereas noise components are discarded. The performance is evaluated in terms of noise removal and feature retention. The comparisons with five state-of-the-art fusion methods and a combination with denoising method demonstrated that our method significantly outperformed the existing techniques with noisy inputs.

We describe a gradient vector flow driven active shape method for model-based image segmentation. Active shape algorithm retain the shape feature of the interested object, and its performance relies heavily on initialization. Because of a lack of global regulation, the control points tends to be trapped in a local optimum in searching. Our proposed method uses the gradient vector flow of an image to guide the optimization process. The control points of an active shape are steered by the direction and the magnitude of gradient vectors. Our experiments demonstrated great improvement in finding the global optimum and resulting correct segmentation.

Niching techniques diversify the population of evolutionary algorithms, encouraging heterogeneous convergence to multiple optima. The key to an effective diversification is identifying the similarity among individuals. With no prior knowledge of the fitness landscapes, it is usually determined by uninformative assumptions on the number of peaks. We propose a method to estimate the sharing distance and the corresponding population size. Using the probably approximately correct (PAC) learning theory and the e-cover concept, we derive the PAC neighbor distance of a local optimum. Within this neighborhood, uniform samples are drawn and we compute the subspace fitness distance correlation (FDC) coefficients. An algorithm is developed to estimate the granularity feature of the fitness landscapes. The sharing distance is determined from the granularity feature and furthermore, the population size is decided. Experiments demonstrate that by using the estimated population size and sharing distance an evolutionary algorithm (EA) correctly identifies multiple optima.

An approach to fuse multiple images based on fuzzy feature representation and evidential reasoning is proposed in this article. Fuzzy set and Dempster-Shafer theory provides a complete framework to describe information naturally and combine weak evidence from multiple sources. Such situations typically arise in the image fusion problems, where a 'real scene' image has to be estimated from incomplete and unreliable observations. By converting images from their spatial domain into the fuzzy evidential representations, decisions are made to aggregate evidence such that a fused image is generated. The proposed fusion approach is evaluated on a broad set of images and promising results are given.

The focus of this paper is the discovery of negative association rules. Such association rules are complementary to the sorts of association rules most often encountered in the literature and have the forms of X/spl rarr/ -Y or -X/spl rarr/Y. We present a rule discovery algorithm that finds a useful subset of valid negative rules. In generating negative rules, we employ a hierarchical graph-structured taxonomy of domain terms. A taxonomy containing classification information records the similarity between items. Given the taxonomy, sibling rules, duplicated from positive rules with a couple of items replaced, are derived together with their estimated confidence. Those sibling rules that bring big confidence deviation are considered candidate negative rules. Our study shows that negative association rules can be discovered efficiently from large database.