Search

Object tracking based on Siamese networks has achieved great success in recent years, but increasingly advanced trackers are also becoming cumbersome, which will severely limit deployment on resource-constrained devices. To solve the above problems, we designed a network with the same or higher tracking performance as other lightweight models based on the SiamFC lightweight tracking model. At the same time, for the problems that the SiamFC tracking network is poor in processing similar semantic information, deformation, illumination change, and scale change, we propose a global attention module and different scale training and testing strategies to solve them. To verify the effectiveness of the proposed algorithm, this paper has done comparative experiments on the ILSVRC, OTB100, VOT2018 datasets. The experimental results show that the method proposed in this paper can significantly improve the performance of the benchmark algorithm. [ABSTRACT FROM AUTHOR]

A variety of open relation extraction systems have been developed in the last decade. And deep learning, especially with attention model, has gained much success in the task of relation classification. Nevertheless, there is, yet, no research reported on classifying open relation tuples to our knowledge. In this paper, we propose a novel semieager learning algorithm (SemiE) to tackle the problem of open relation classification. Different from the eager learning approaches (e.g., ANNs) and the lazy learning approaches (e.g., kNN), the SemiE offers the benefits of both categories of learning scheme, with its significantly lower computational cost (O(n)). This algorithm can also be employed in other classification tasks. Additionally, this paper presents an adapted attention model to transform relation phrases into vectors by using word embedding. The experimental results on two benchmark datasets show that our method outperforms the state-of-the-art methods in the task of open relation classification. [ABSTRACT FROM AUTHOR]

Person reidentification is aimed at solving the problem of matching and identifying people under the scene of cross cameras. However, due to the complicated changes of different surveillance scenes, the error rate of person reidentification exists greatly. In order to solve this problem and improve the accuracy of person reidentification, a new method is proposed, which is integrated by attention mechanism, hard sample acceleration, and similarity optimization. First, the bilinear channel fusion attention mechanism is introduced to improve the bottleneck of ResNet50 and fine-grained information in the way of multireceptive field feature channel fusion is fully learnt, which enhances the robustness of pedestrian features. Meanwhile, a hard sample selection mechanism is designed on the basis of the P2G optimization model, which can simplify and accelerate picking out hard samples. The hard samples are used as the objects of similarity optimization to realize the compression of the model and the enhancement of the generalization ability. Finally, a local and global feature similarity fusion module is designed, in which the weights of each part are learned through the training process, and the importance of key parts is automatically perceived. Experimental results on Market-1501 and CUHK03 datasets show that, compared with existing methods, the algorithm in this paper can effectively improve the accuracy of person reidentification. [ABSTRACT FROM AUTHOR]

The registration of point clouds in a three-dimensional space is an important task in many areas of computer vision, including robotics and autonomous driving. The purpose of registration is to find a rigid geometric transformation to align two point clouds. The registration problem can be affected by noise and partiality (two point clouds only have a partial overlap). The Iterative Closed Point (ICP) algorithm is a common method for solving the registration problem. Recently, artificial neural networks have begun to be used in the registration of point clouds. The drawback of ICP and other registration algorithms is the possible convergence to a local minimum. Thus, an important characteristic of a registration algorithm is the ability to avoid local minima. In this paper, we propose an ICP-type registration algorithm (λ -ICP) that uses a multiparameter functional (λ -functional). The proposed λ -ICP algorithm generalizes the NICP algorithm (normal ICP). The application of the λ -functional requires a consistent choice of the eigenvectors of the covariance matrix of two point clouds. The paper also proposes an algorithm for choosing the directions of eigenvectors. The performance of the proposed λ -ICP algorithm is compared with that of a standard point-to-point ICP and neural network Deep Closest Points (DCP). [ABSTRACT FROM AUTHOR]

We construct a neural network based on smoothing approximation techniques and projected gradient method to solve a kind of sparse reconstruction problems. Neural network can be implemented by circuits and can be seen as an importantmethod for solving optimization problems, especially large scale problems. Smoothing approximation is an efficient technique for solving nonsmooth optimization problems. We combine these two techniques to overcome the difficulties of the choices of the step size in discrete algorithms and the item in the set-valued map of differential inclusion. In theory, the proposed network can converge to the optimal solution set of the given problem. Furthermore, some numerical experiments show the effectiveness of the proposed network in this paper. [ABSTRACT FROM AUTHOR]

This paper investigates the problem of impact-time-control and proposes a learning-based computational guidance algorithm to solve this problem. The proposed guidance algorithm is developed based on a general prediction-correction concept: the exact time-to-go under proportional navigation guidance with realistic aerodynamic characteristics is estimated by a deep neural network and a biased command to nullify the impact time error is developed by utilizing the emerging reinforcement learning techniques. To deal with the problem of insufficient training data, a transfer-ensemble learning approach is proposed to train the deep neural network. The deep neural network is augmented into the reinforcement learning block to resolve the issue of sparse reward that has been observed in typical reinforcement learning formulation. Extensive numerical simulations are conducted to support the proposed algorithm. [ABSTRACT FROM AUTHOR]

As an essential subproblem in the face recognition field, occluded face recognition has received considerable attention in recent years. However, satisfactory recognition accuracy for occluded faces has not been achieved. Therefore, in order to solve this problem, this paper proposes a convolutional neural network that has multidimensional serial feature extraction modules for occluded faces and uses the deep learning method to improve the recognition rate. In order to improve the expression ability of the network, the method extracts features from two dimensions — space and channel — and built a multidimensional serial feature extraction module. By "multiscale" and "dependence" processing features, the generalization ability was significantly improved. Through the experimental test, the recognition accuracy of the multidimensional feature network (MFNet) reached 90.35%. This accuracy is 17.89% (72.36%) better than that of the traditional algorithm. Meanwhile, compared with other convolutional neural networks, this network has different degrees of improvement, which are 0.68% (ArcFace), 2.14% (Visual Geometry Group Face), and 4.83% (DeepID3). [ABSTRACT FROM AUTHOR]

Abstract: Skin color detection is a popular and useful technique because of its wide range of utilizations both in human computer interaction and content based analysis. Applications such as detecting and tracking of human body parts, face detection and recognition, naked people detection and people retrieval in multimedia databases all benefit from skin detection. Hence finding a proper method for segment the skin-like pixels can solve the presented problems. The Imperialist Competitive Algorithm (ICA) is a new evolutionary algorithm which was recently introduced and has a good performance in some optimization problems. The ICA was inspired by socio-political processes of imperialistic competition of mankind in the real world. In this paper a hybrid ICA-ANN is proposed for solving skin classification. A new algorithm that combines ICA and ANN to solve skin classification has been proposed, and then a hybrid Neural Network (NN)-Imperialist Competitive Algorithm (ICA) is applied to solve the classification problem. In the proposed algorithm, a multi layer perceptron network (MLP) manages the problem’s constraints and an ICA algorithm searches for high quality solutions and minimum cost. The proposed color segmentation algorithm operates directly on RGB color space without the need of color space conversion. Experimental results show that this method can improve the performance of the MLP algorithm significantly. [Copyright &y& Elsevier]

Abstract: For nonlinear processes the classical model predictive control (MPC) algorithm, in which a linear model is used, usually does not give satisfactory closed-loop performance. In such nonlinear cases a suboptimal MPC strategy is typically used in which the nonlinear model is successively linearised on-line for the current operating point and, thanks to linearisation, the control policy is calculated from a quadratic programming problem. Although the suboptimal MPC algorithm frequently gives good results, for some nonlinear processes it would be beneficial to further improve control accuracy. This paper details a computationally efficient nonlinear MPC algorithm in which a neural model is linearised on-line along the predicted trajectory in an iterative way. The algorithm needs solving on-line only a series of quadratic programming problems. Advantages of the discussed algorithm are demonstrated in the control system of a high-purity ethylene–ethane distillation column for which the classical linear MPC algorithm does not work and the classical suboptimal MPC algorithm is slow. It is shown that the discussed algorithm can give practically the same control accuracy as the algorithm with on-line nonlinear optimisation and, at the same time, the algorithm is significantly less computationally demanding. [Copyright &y& Elsevier]

Revealing the corresponding identities of the same individual in different systems is a common task in various areas, e.g., criminals inter-network tracking, homologous proteins revealing, ancient words translating, and so on. With the reason that, recently, more and more complex systems are described by networks, this task can also be accomplished by solving a node matching problem among these networks. Revealing one-to-one matching between networks is for sure the best if we can, however, when the target networks are highly symmetric, or an individual has different identities (corresponds to several nodes) in the same network, the exact one-to-one node matching algorithms always lose their effects to obtain acceptable results. In such situations, one-to-many (or many-to-many) node matching algorithms may be more useful. In this paper, we propose two one-to-many node matching algorithms based on local mapping and ensembling, respectively. Although such algorithms may not tell us the exact correspondence of the identities in different systems, they can indeed help us to narrow down the inter-network searching range, and thus are of significance in practical applications. These results have been verified by the matching experiments on pairwise artificial networks and real-world networks. [ABSTRACT FROM AUTHOR]

In this paper, we propose two cooperative ensemble learning algorithms, i.e., NegBagg and NegBoost, for designing neural network (NN) ensembles. The proposed algorithms incrementally train different individual NNs in an ensemble using the negative correlation learning algorithm. Bagging and boosting algorithms are used in NegBagg and NegBoost, respectively, to create different training sets for different NNs in the ensemble. The idea behind using negative correlation learning in conjunction with the bagging/boosting algorithm is to facilitate interaction and cooperation among NNs during their training. Both NegBagg and NegBoost use a constructive approach to automatically determine the number of hidden neurons for NNs. NegBoost also uses the constructive approach to automatically determine the number of NNs for the ensemble. The two algorithms have been tested on a number of benchmark problems in machine learning and NNs, including Australian credit card assessment, breast cancer, diabetes, glass, heart disease, letter recognition, satellite, soybean, and waveform problems. The experimental results show that NegBagg and NegBoost require a small number of training epochs to produce compact NN ensembles with good generalization. [ABSTRACT FROM AUTHOR]

Support vector machines (SVMs) have been extensively used. However, it is known that SVMs face difficulty in solving large complex problems due to the intensive computation involved in their training algorithms, which are at least quadratic with respect to the number of training examples. This paper proposes a new, simple, and efficient network architecture which consists of several SVMs each trained on a small subregion of the whole data sampling space and the same number of simple neural quantizer modules which inhibit the outputs of all the remote SVMs and only allow a single local SVM to fire (produce actual output) at any time. In principle, this region-computing based modular network method can significantly reduce the learning time of SYM algorithms without sacrificing much generalization performance. The experiments on a few real large complex bench- mark problems demonstrate that our method can be significantly faster than single SVMs without losing much generalization performance. [ABSTRACT FROM AUTHOR]

In this paper, we present a method for retrieving sea surface wind field (SSWF) from HaiYang-2B (HY-2B) scatterometer data. In contrast to the conventional algorithm, i.e., using a point-to-point (P2P) method based on geophysical model functions (GMF) to retrieve SSWF by spaceborne scatterometer, we introduce a more accurate field-to-field (F2F) retrieval method based on convolutional neural network (CNN). We fully consider the spatial correlation and continuity between adjacent observation points, and input the observation data of continuous wind field within a certain range into the neural network to construct the neural network model, and then synchronously obtain the wind field within the range. The wind field obtained by our retrieval method maintains its continuity and solves the problem of ambiguity removal in traditional wind direction retrieval methods. Comparing the retrieval results with the buoy data, the results show that the root mean square errors (RMSE) of wind direction and wind speed are less than 0.18 rad (10.31°) and 0.75 m/s, respectively. The retrieval accuracy is better than the L2B product of HY-2B. [ABSTRACT FROM AUTHOR]

In this paper, a proposed algorithm that dynamically changes the neural network structure is presented. The structure is changed based on some features in the cascade correlation algorithm. Cascade correlation is an important algorithm that is used to solve the actual problem by artificial neural networks as a new architecture and supervised learning algorithm. This process optimizes the architectures of the network which intends to accelerate the learning process and produce better performance in generalization. Many researchers have to date proposed several growing algorithms to optimize the feedforward neural network architectures. The proposed algorithm has been tested on various medical data sets. The results prove that the proposed algorithm is a better method to evaluate the accuracy and flexibility resulting from it. [ABSTRACT FROM AUTHOR]