Despite progress in 3D object recognition using deep learning (DL), challenges such as domain shift, occlusion, and viewpoint variations hinder robust performance. Additionally, the high computational cost and lack of labeled data limit real-time deployment in applications such as autonomous driving and robotic manipulation. To address these challenges, we propose 3D-CDNeT, a novel cross-domain deep learning network designed for unsupervised learning, enabling efficient and robust point cloud recognition. At the core of our model is a lightweight graph-infused attention encoder (GIAE) that enables effective feature interaction between the source and target domains. It not only improves recognition accuracy but also reduces inference time, which is essential for real-time applications. To enhance robustness and adaptability, we introduce a feature invariance learning module (FILM) using contrastive loss for learning invariant features. In addition, we adopt a Generative Decoder (GD) based on a Variational Auto-Encoder (VAE) to model diverse latent spaces and reconstruct meaningful 3D structures from the point cloud. This reconstruction process acts as a self-supervised generative objective that complements the discriminative recognition task, guiding the encoder to learn structure-preserving and domain-invariant features that improve recognition under occlusion and cross-domain conditions. Our proposed model unifies generative and discriminative tasks by using self-attention on the object covariance matrix to facilitate efficient information exchange, enabling the extraction of both local and global features. We further develop a self-supervised pretraining strategy that learns both global and local object invariances through GIAE and GD, respectively. A new loss function, combining contrastive loss and Chamfer distance, is proposed to strengthen cross-domain feature alignment. Experimental results on three benchmark datasets demonstrate that 3D-CDNeT outperforms existing state-of-the-art (SOTA) methods in recognition accuracy and inference speed, offering a practical solution for real-time 3D perception tasks. It achieves accuracies of 90.6 % on ModelNet40, 95.2 % on ModelNet10, and 76.4 % on the ScanObjectNN dataset in linear evaluation tasks, all while reducing runtime by 45 % without compromising performance. Detailed qualitative comparisons and ablation studies are provided to validate the effectiveness of each component and demonstrate the superior performance of our proposed method.

Learning-based multi-view stereo has made significant advances. However, some methods neglect the guidance of various geometric priors, which leads to poor reconstruction under challenges like occlusions and weak textures. Meanwhile, other works only explore the contribution of a single geometric prior at a relatively fixed stage, which lacks holistic constraints and complementary collaboration. Therefore, this paper integrates surface normals, epipolar geometry, visibility, and geometric consistency through the entire pipeline. It aims to resolve various issues through the collaboration of multiple complementary geometric cues at different stages. The Key innovations include a dual guidance of surface normals in feature extraction and volume aggregation, which enhances the network’s robustness to occlusion through geometric visibility and epipolar-enhanced feature group correlation, while also reinforcing geometric structural features. Along with a frequency-optimized consistency inpainting pipeline through multi-direction and multi-stage. With the further utilization of depth optimization modules, extensive experiments on 5 datasets of varying scales demonstrate that our proposed MGPG-MVSNet exhibits excellent generalization to real-world engineering applications, particularly achieving superior reconstruction in occluded and textureless regions.

Three-dimensional (3D) human pose estimation garners increasing attention in industrial applications, particularly in intelligent healthcare, sports analytics, and VR/AR systems. While Transformer-based methods achieve significant progress, they suffer from quadratic computational complexity and limited long-sequence modeling capacity. Pure State Space Model (SSM) approaches offer linear complexity but lack explicit spatial relationship modeling. We propose SCTMPose, a skeleton-constrained Transformer-Mamba hybrid architecture for 3D human pose estimation. Our key innovation employs a hybrid integration strategy that replaces the static Feed-Forward Networks (FFNs) of traditional Transformers with dynamic temporal Mamba blocks. This design achieves superior computational efficiency by leveraging the linear complexity of SSMs while preserving spatiotemporal relationship modeling strengths. Furthermore, we propose a Skeletal Constraint Module (SCM) that explicitly models biological constraints through spatial multi-head attention and graph-based message passing with learnable adjacency matrices. Through comprehensive exploration of Transformer-Mamba integration strategies, we establish optimal design principles for hybrid spatiotemporal architectures. Extensive experiments demonstrate that SCTMPose achieves state-of-the-art (SOTA) performance with 37.8 mm MPJPE on Human3.6M and 13.9 mm MPJPE on MPI-INF-3DHP, delivering significant improvements in accuracy, efficiency, and biological plausibility.1.

Pedestrian trajectory prediction and behavior analysis are crucial in intelligent transportation systems (ITS), and the key issue is to effectively achieve pedestrian re-identification (ReID) from cross-camera networks. But however, clothing changes can greatly affect the accuracy of recognition. Since clothing and identity have a complex relationship, existing models inadequately separate identity features from clothing-induced bias through causal analysis. Moreover, current multi-modality methods often overlook the descriptive attributes of individuals in the original RGB images. Therefore, this paper proposes a Causal Textual Visual Network (CTVNet) for Cloth-changing ReID. CTVNet comprises three branches: clothing, identity, and a parallel textual branch, with the textual branch being the primary contribution. This branch introduces two novel modules including an Attribute Extraction and Masking (AEM) module and a Multi-modality Coordination Network (MCN) to unify attribute descriptions with the original RGB images. The parallel textual branch extracts text descriptions unrelated to clothing, fuses them with visual features, and subtracts from identity features to eliminate redundant clothing information, enabling causal intervention. The AEM module masks color and clothing information in descriptive attributes, while the MCN integrates features across tokens and channels, leveraging their effectiveness. For the first time, causal intervention is combined with textual attributes in multi-modality cloth-changing ReID (CC-ReID). Furthermore, masked attribute descriptions are combined with visual features fused by the MCN to reduce the influence of clothing and eliminate residual clothing bias through causal intervention. Extensive experiments on two standard CC-ReID datasets demonstrate the superiority of the proposed CTVNet. Both qualitative and quantitative results are provided. Additionally, several ablation studies are conducted on each component of the proposed model to demonstrate their effectiveness.

Considering the characteristics of Internet-of-Things enabled multi-robot maritime patrolling problems which are not reflected in general multi-objective optimization benchmarks, this work proposes a similarity evaluation framework to enhance the diversity when the population evolves. A novel masking thinking is proposed to eliminate the effects of dominant and highly correlated genetic positions and thus magnify the impacts of the rest which can distinguish individuals more effectively. The perspective from objective space is also considered to maintain a desired balance between diversity and convergence. With the use of the proposed similarity evaluation framework, individuals contributing more to the diversity can be selected to the next generation and further evolved. It is beneficial to prevent a population from being trapped by local optima and thus explore better Pareto-optimal solutions. The effectiveness of the proposed framework is validated by ablation experiments. Comparisons to the state of the art are also conducted to illustrate its advantages.

Multirobot patrolling systems with various sensing and communications devices are deployed to guarantee maritime safety. Patrolling path planning for multiple robots can be modeled as a multiobjective optimization problem. The positions of patrolling nodes impact the length of patrolling paths and execution efficiency of robots. To compute them, a huge solution space is encountered. Besides, multiple patrolling nodes on the same line lead to the same patrolling scheme. Thus, how to promote solution (population) diversity becomes a new challenge. To tackle it, this work proposes a learning-inspired immune algorithm. It uses the historical information in the previous generations during iterations to realize a learning process. Unlike saving all the individuals themselves and training a model for them, the useful historical information is extracted by using upper confidence bound-based and actor–critic-inspired methods. Both time consumption and storage space can be dramatically saved. The experimental results indicate that the proposed algorithm can generate multiple patrolling schemes for the decision makers and outperforms the state-of-the-art.

The seven most common skin diseases are melanocytic nevus, melanoma, benign keratosis, basal cell carcinoma, actinic keratosis, vascular lesions, and dermatofibroma. Among them, melanoma has been identified as one of the deadliest cancers based on medical studies and research. The current trend in disease detection revolves around the use of machine learning and deep learning models. Regardless of the model used, the crucial aspect is achieving accurate classification for these diseases. With the emergence of powerful convolutional neural networks (CNNs), significant progress has been made in classification of skin cancer lesions in recent years. However, various challenges hinder the development of practical and effective solutions. First, due to the specific nature of skin cancer lesion images, the current deep neural network architectures and training strategies have poor adaptability to medical images. They are also prone to gradient vanishing issues during network iteration, which hinders the construction of high-performance deep learning models that leverage distinctive characteristics of skin lesion images. Second, there exists a discordance between skin lesion images and deep learning network structures. To address these issues, this study introduces a soft attention mechanism to enhance adaptability to skin cancer lesion images and improve the extraction of informative features from medical images. Additionally, a novel multi-scale fusion convolutional neural network model is proposed to overcome the mismatch between deep learning CNN architectures and skin lesion images. This model autonomously extracts appearance features from raw dermatological medical images. Comparisons with other popular techniques demonstrate the effectiveness of the proposed model, which can achieve an accuracy of 93.9% on HAM10000 dataset. There is ongoing research to overcome the remaining challenges and further enhance the performance of skin cancer classification algorithms.

This paper proposes a few-shot pedestrian re-identification (Re-ID) model based on an improved ResNet50 with a compression and stimulation module, which is named CS-ResNet50. It combines the meta-learning framework with metric learning. This method first compresses residual network channels, then stimulates them to achieve the effect of feature weighting, ultimately making feature extraction more accurate. The research makes the model learn how to finish new tasks efficiently from its experience that it has obtained in the training process of former subtasks. In each subtask, the dataset is divided into a gallery set and a query set, where the model parameters are trained. In this way, the model can be trained efficiently and adopted to new tasks rapidly, which could solve few-shot Re-ID problems. Compared with the baseline, the proposed model improves two indicators efficiently on two Re-ID datasets and achieves better Re-ID effect in few-shot mode.

In multirobot patrolling problems, various dynamic situations require a higher level of cooperation among robots. The dynamic problems caused by moving objects are rarely studied yet. This work proposes a distributed event-driven cooperative strategy for multirobot systems to patrol moving objects autonomously. First, forward and backward utility functions are defined as criteria for robots to conduct two-way evaluation when they choose their targets to patrol. Then, three event types and a cooperative action considering energy consumption and visiting frequency comprehensively are proposed to improve coordination among robots during their execution processes. In simulation experiments, the proposed strategy shows significant advantages on decreasing the average and maximum unvisited time of moving objects compared with the state-of-the-art. A marine pollution monitoring case is simulated to demonstrate the practicability of this strategy.

Visible–infrared person re-identification (VI-ReID) is a current focused area in the field of re-identification. In order to reduce the gap between two modalities in VI-ReID and improve recognition accuracy, this paper proposes a four-stream network and nonsignificant feature learning (FS-NSF) method for VI-ReID. First, the dual-intermediate modality images of visible and infrared modalities are generated by two lightweight networks, and the labels are inherited from the visible and infrared images. Second, the ResNet50 backbone network is split in order to reconstruct the network adapted to shared feature learning of the four modalities. Finally, a multi-branch, multi-scale and multi-granularity feature extraction strategy is used to extract both significant and nonsignificant features. The comparison experiments are conducted on SYSU-MM01 dataset and RegDB dataset. The experimental results show that, compared with state-of-the-arts, our method has excellent performance on both datasets, especially on the SYSU-MM01 dataset, with an increase in performance of 1.9–6.28% for each index.

Rate-Distortion Optimization (RDO) is an important step in video coding to achieve the best quality under a certain compression ratio constraint. The traditional RDO assigns equal importance to different color components. However, Human Visual System (HVS) has different sensitivities to different components. In this paper, the color-sensitivity-based combined PSNR (CSPSNR) is utilized as the distortion measurement in the process of RDO, where the characteristics of the color sensitivities of HVS are taken into account. Firstly, the distortion weights of luma and chroma components are derived from the criterion of maximizing CSPSNR. Then Lagrange multiplier and quantization parameter (QP) are adjusted according to the variation of distortion weights among different components. Finally, the CSPSNR-based RDO (CSRDO) adaptively calculates the RD costs of luma and chroma components under different sampling rates to improve the coding efficiency of the whole sequence. Experimental results in H.265/HEVC demonstrate that the proposed method can achieve 3.11% and 3.58% BD-RATE gain for AI and RA configurations in terms of CSPSNR on average.

With the continuous improvement and development of cameras network, surveillance video has become the data source of the column stream, which greatly promotes the development of cross-camera person re-identification (Re-ID). However, supervised learning requires a lot of effort to manually label cross-cameras pairwise training data, which is lack of scalability and practical in actual video surveillance because there is a lack of well-labeled pairs of positive and negative samples under each camera. For addressing these negative effects, we set judgment conditions by using the association ranking method to self-discover positive and negative track-lets pairs of anchors with none of the pairwise ID labels, thereby defining a triplet loss. In order to optimize association loss for learning effective discriminative feature, the triplet loss adds adaptive weights according to the degree of easy-hard samples to generate an Adaptive Weighted Conditional Triplet Loss. Besides, for increasing the accuracy of self-discovering cross-camera anchors independently, which means successfully mine mutually best-matched track-lets and merge them under cross-camera, we use the top-rank from the intra-camera ranking list as a self-matched query sample which can double verify the matched-degree between top-rank. And eventually, we establish a new Association Loss and Self-Discovery Learning (ALSL) model with a complete end-to-end manner. We use three standard datasets, PRID2011, iLIDS-VID and MARS, to train the model and the experimental results prove that ALSL rank-1 is better than some superior video-based unsupervised person Re-ID methods.

One of the important Internet-of-Things applications is to use image and video to realize automatic people monitoring, surveillance, tracking, and reidentification (Re-ID). Despite some recent advances, pedestrian Re-ID remains a challenging task. Existing algorithms based on fully supervised learning for it usually requires numerous labeled image and video data, while often ignoring the problem of data imbalance. This work proposes a method based on unlabeled samples generated by cycle generative adversarial networks. For a newly generated unlabeled sample, it learns its pseudorelationship between unlabeled samples and labeled ones in a low-dimensional space by using a self-paced learning approach. Then, these unlabeled ones having pseudo-relationship with labeled ones are added in a training set to better mine discriminative information between positive and negative samples, which is in turn used to learn a more effective metric. We name this method as a semi-supervised learning approach based on the built pseudopairwise relations between labeled data and unlabeled one. It can greatly enhance the performance of pedestrian Re-ID in case of insufficient labeled images. By using only about 10% labeled images in a given database, the proposed method obtains higher accuracy than state-of-the-art supervised learning methods using all labeled ones, e.g., deep-learning ones, thus greatly advancing the field of pedestrian Re-ID.

Pedestrian re-identification (Re-ID) is a very challenging and unavoidable problem in the field of multi-camera surveillance in smart transportation. Among many ways to solve this problem, keep it simple and straightforward (KISS) metric learning (KISSME) stands out since it has unbeatable advantages in running time while maintaining highly acceptable matching rate. It can be used to realize effective pedestrian Re-ID in an open world. Although it has achieved highly acceptable performance in some applications, it encounters a small sample size (S3) problem that causes too small eigenvalues of its covariance matrix, thus resulting in an instability issue. Its large eigenvalues are overestimated; while its small ones are underestimated. In order to solve this problem, we use an orthogonal basis vector to generate virtual samples to overcome the S3problem. The resulting algorithm named KISS+ is experimentally shown to have the eigenvalues of its covariance matrix significantly larger than those of the original KISSME. In order to show its advantage in pedestrian Re-ID, this work uses multi-feature fusion to extract more discriminant features, and obtain a low-dimensional expression of features through dimension reduction. Experiments based on several well-known databases show that our method can improve the matching rate, while maintaining the advantage of fast computation. Compared with deep learning algorithms, our algorithm does not achieve their matching rate, but it is highly suitable for real-time pedestrian Re-ID of an open world due to its simplicity, easy operation and fast execution.

Person re-identification (person re-ID) is a challenging task which aims at spotting same persons among disjoint camera views. It has certainly generated a lot of attention in the field of computer vision, but it remains a challenging task due to the complexity of person appearances from different camera views. To solve this challenging problem, many excellent methods have been proposed, especially metric learning-based algorithms. However, most of them suffer from the problem of data imbalance. To solve this problem, in the paper we proposed a new data-balanced method and named it Enhanced Metric Learning (EML) based on adaptive asymmetric and diversity regularization for person re-ID. Metric learning is important for person re-ID because it can eliminate the negative effects caused by camera differences to a certain extent. But most metric learning approaches often neglect the problem of data imbalance caused by too many negative samples but few positive samples. And they often treat all negative samples the same as positive ones, which can lead to the loss of important information. Our approach pays different attention to the positive samples and negative ones. Firstly, we classified negative samples into three groups adaptively, and then paid different attention to them using adaptive asymmetric strategy. By treating samples differently, the proposed method can better exploit the discriminative information between positive and negative samples. Furthermore, we also proposed to impose a diversity regularizer to avoid over-fitting when the training sets are small or medium-sized. Finally, we designed a series of experiments on four challenging databases (VIPeR, PRID450S, CUHK01 and GRID), to compare with some excellent metric learning methods. Experimental results show that the rank-1 matching rate of the proposed method has outperformed the state-of-the-art by 3.64%, 4.2%, 3.13% and 2.83% on the four databases, respectively.

Person re-identification (Re-ID) is a research hot spot in the field of intelligent video analysis, and it is also a challenging task. As the number of samples grows larger, traditional metric and feature learning methods fall into bottleneck, while it just meets the needs of deep learning algorithm, which perform very well in person re-identification. Although they have achieved good results in the field of supervised learning, their application in real-world scenarios is not very satisfactory. This is mainly because in the real world, a huge number of labeled images are hard to obtain, and even if they are obtained, the cost is expensive. Meanwhile, the performance of deep learning in unsupervised metrics is not ideal. For solving the problem, we propose a new method based on unsupervised domain adaptation (UDA) and re-ranking, and name it UDA[Formula: see text]. As for this method, we first train a camera-aware style transfer model to gain camstyle images. Then we further reduce the difference between the domain of the target and source by using invariant feature, and further improve their commonality. In addition, re-ranking is also introduced to optimize the matching results. This method can not only reduce the cost of obtaining labeled data, but also improve the accuracy. Experimental results show that our method can outperform the most advanced method by 4% on Rank-1 and 14% on mAP. The results also better confirm the effectiveness of Re-ranking module and provide a new idea for domain adaptation by unsupervised methods in the future.

This work investigates whether virtual samples can solve the small sample size (S3) problem of Keep-It-Simple-and-Straightforward Metric Learning (KISSME) in Pedestrian re-identification (Re-ID). Re-ID is a very challenging and important problem in the field of multi-camera surveillance in smart transportation. Among many hand-crafted ways (not deeply-learned ones) to solve it, KISSME has received great attention. Although it has achieved convincing performance in some applications, it encounters an S3problem in calculating various classes of covariance matrices whose eigenvalues become too small. Such small eigenvalues cause an instability issue when computing the inverse of covariance matrices, thus resulting in poor Re-ID performance. If we can increase the number of samples, then an S3problem is alleviated or eliminated. This work makes a hypothesis that virtual samples can do so, and proposes a new algorithm to generate them. It adopts a Genetic Algorithm to generate virtual features (corresponding to virtual samples) based on the dimension-reduced sample features, which eliminates the process of re-extracting features of newly generated virtual samples and save time. It can clearly increase the magnitude of otherwise small eigenvalues, helps one perform the accurate estimation of the inverse of various covariance matrices and finally alleviates the S3problem. Experimental results based on a commonly-used database confirm that the proposed method can significantly improve the matching rate of pedestrian Re-ID, which fully shows that virtual samples are indeed effective for alleviating the S3problem in pedestrian Re-ID.

Person re-identification via cross-camera is a difficult problem in the field of target discovery and tracking. Traditional solutions depending on the characteristics of a target's appearance have low reliability and can easily lead to low matching rate because they use simple metric functions. This work proposes a more reliable measurement: correlative degree of a target's features among different camera views to do person re-identification and uses it to measure the histograms' similarity of targets. In order to obtain more discriminative features, we need to extract their appearance and space features. We use multi-layer histograms to describe them. In order to compute more accurate correlation degree, we propose to use Gaussian pyramid as alternating distance to define high-dimensional diffusion distance. Finally, we assign different weights to feature vectors so as to establish the correlative degree function based on diffusion distance. Experiments of person re-identification for different cameras show that the proposed method can achieve much better results than some known existing methods.

In the wireless sensor networks with multiple mobile sinks, the movement of sinks or failure of sensor nodes may leads to the breakage of existing routes. In order to repair broken path with lower communication overhead in terms of both energy and delay, we propose an efficient routing recovery protocol with endocrine cooperative particle swarm optimization algorithm to establish and optimize the alternative path. With this method, the alternative path from source nodes to the sink with the optimal QoS parameters can be selected. Simulation results demonstrate that ECPSOA can adapt to rapid topological changes with multiple mobile sinks, while decreasing communication overhead and efficiently reducing the energy consumption.