Spectral clustering is a prevalent clustering method in which an affinity matrix is constructed based on all samples (the number is n ), leading to high computational complexity and making it infeasible for dealing with large-scale data directly. In this study, we introduce an Approximate Spectral Embedding Representation method (ASER). By employing an anchor-based strategy, the spectral embedding representation of the selected anchors (the number is m , m ≪ n ) is used to approximate the spectral embedding representation of the original samples. Unlike available methods that approximate the similarity matrix based on an anchor graph, we directly implement the approximation in the spectral embedding space. Moreover, the properties of the formed anchor graph are inherited from the original space to the spectral embedding space. The time complexity of conducting spectral clustering is significantly reduced from O ( n 3 ) to be linear with respect to n , without relying on any acceleration operations for eigenvalue decomposition. Experimental results on toy examples and benchmark datasets with large sizes demonstrate the effectiveness and efficiency of the proposed model.

The HEVC standard, designed to outperform H.264/AVC in compression efficiency, incorporates several advanced techniques in intra coding, including coding tree units (CTUs) and an expanded set of 35 prediction modes per prediction unit (PU). While these innovations enhance coding performance, they also escalate computational complexity. We present a fast intra-mode decision method for HEVC to lower complexity. First, the algorithm integrates the grid design characteristics of intra-prediction, tailoring texture gradient prediction angles for coding units (CUs) of varying sizes. Second, a dynamic preselection method leveraging the cumulative energy of texture gradients and the Most Probable Mode (MPM) mechanism is employed to efficiently filter candidate prediction modes. Experimental results show the proposed algorithm improves coding efficiency by$\mathbf{2 2. 3 \%}$over HM18.0, with only$\mathbf{1. 6 \%}$average bitrate increase.

Night time vehicle detection and tracking using traditional visible light cameras is a challenging task due to the limited visibility. The current state-of-the-art systems treat the vehicles at night time as paired vehicle headlights or taillights, with no ability to determine the contour of the vehicle or its spatial occupancy. Therefore, this paper proposes the first night time framework that combines the vehicle headlights and taillights to localize the vehicle contours. This new framework includes a novel multi-camera vehicle representation that groups and reconstructs vehicle headlights and taillights following mutual geometric distances between different vehicle components. This novel vehicle contour representation successfully removes duplicated vehicle lights and also compensates for the missing vehicle lights in the detection process. Eventually, vehicle headlight alignment and contour adjustment are used to further refine the vehicle contours. The proposed multi-camera system considers typical four-wheel vehicles, e.g., cars and SUVs, in the monitoring and might not be able to handle large trucks, e.g., 18-wheelers. The experiments are conducted on night time traffic videos under various scenarios and the proposed system attains an average of 0.896 in Multiple Object Tracking Accuracy (MOTA) and an average of 0.904 in Jaccard Coefficient (JC), which indicates 19.2% and 15.9% increases over the baseline system. The night time traffic datasets used in this paper are available athttps://github.com/JustMeZXX/Intelligent-Night-Time-Traffic-Surveillance

In heterogeneous cellular network, task scheduling for computation offloading is one of the biggest challenges. Most works focus on alleviating heavy burden of macro base stations by moving the computation tasks on macro cell user equipment (MUE) to remote cloud or small cell base stations. But the selfishness of network users is seldom considered. Motivated by the multiple access mobile edge computing, this paper provides incentive for task transfer from macro cell users to small cell base stations. The proposed incentive scheme utilizes small cell user equipments to provide relay services. The problem of computation offloading is modeled as a two-stage auction, in which the remote MUEs with common social character can form a group and then buy the computation resource of small cell base stations with the relaying of small cell user equipment. A two-stage auction scheme named TARCO is contributed to maximize utilities for both sellers and buyers in the network. The truthfulness, individual rational and budget balance properties of TARCO are also proved in this paper. In addition, two algorithms are proposed to further refine TARCO on the social welfare of the network. One can achieve higher utility of MUEs and the other can obtain higher total social welfare. Extensive simulation results demonstrate that, TARCO is better than random algorithm by 104.90 percent in terms of average utility of MUEs, while the performance of TARCO is further improved up to 28.75 percent and 17.06 percent by the proposed two algorithms, respectively.

Night time vehicle detection and tracking has been a challenging task in recent years. This paper presents a novel context-aware traffic surveillance system that integrates sensor information from autonomous vehicles to improve performance of night time vehicle detection and tracking. The key elements of the proposed method include a novel vehicle pairing framework that represents vehicles based on the fused sensor contexts and vehicle taillights. These detected vehicles are then tracked in real-time night time traffic videos. Experiments are conducted on real traffic videos and the proposed system attains 0.6319 in multiple object tracking accuracy (MOTA), which represents a 26.1% increase compared with the baseline performance.

Long Short-Term Memory (LSTM) has been proven an efficient way to model sequential data, because of its ability to overcome the gradient diminishing problem during training. However, due to the limited memory capacity in LSTM cells, LSTM is weak in capturing long-time dependency in sequential data. To address this challenge, we propose an Attention-aware Bidirectional Multi-residual Recurrent Neural Network (ABMRNN) to overcome the deficiency. Our model considers both past and future information at every time step with omniscient attention based on LSTM. In addition to that, the multi-residual mechanism has been leveraged in our model which aims to model the relationship between current time step with further distant time steps instead of a just previous time step. The results of experiments show that our model achieves state-of-the-art performance in classification tasks.

Computer-vision methods have been extensively used in intelligent transportation systems for vehicle detection. However, the detection of severely occluded or partially observed vehicles due to the limited camera fields of view remains a challenge. This paper presents a multi-camera vehicle detection system that significantly improves the detection performance under occlusion conditions. The key elements of the proposed method include a novel multi-view region proposal network that localizes the candidate vehicles on the ground plane. We also infer the vehicle position on the ground plane by leveraging multi-view cross-camera context. Experiments are conducted on dataset captured from a roadway in Richardson, TX, USA, and the system attains 0.7849 Average Precision and 0.7089 Multi Object Detection Precision. The proposed system results in an approximately 31.2% increase in AP and 8.6% in MODP than the single-camera methods.

Unlike previous unknown nouns tagging task, this is the first attempt to focus on out-of-vocabulary (OOV) lexical evaluation tasks that does not require any prior knowledge. The OOV words are words that only appear in test samples. The goal of tasks is to provide solutions for OOV lexical classification and predication. The tasks require annotators to conclude the attributes of the OOV words based on their related contexts. Then, we utilize unsupervised word embedding methods such as Word2Vec and Word2GM to perform the baseline experiments on the categorical classification task and OOV words attribute prediction tasks.

Vision-based techniques are gradually being applied to civil engineering fields by aiding human resources to inspect and assess the condition of structures. This paper presents a vision-based inspection system to detect bends of the bridge chords in UAV images. The main parts of this paper firstly present a novel region proposal algorithm to localize bridge chord bends without distinct boundaries and then present a fusion-based deep semantic segmentation network that improves the detection performance by fusing spatial information from neighborhood. The proposed system achieves approximate 0.7762 Mean IOU and 0.6424 BF Score on the field dataset, which outperforms the other state-of-the-art systems.

our task of video copy detection system aims to locate vicdeo segments that are partially copied or near-duplicated versions from an archive of reference videos. In 2010, video copy detection problem was sometimes considered as a solved problem, since previous research within this area used either small-scale or large-scale datasets (e.g. TRECVID 2009, Muscle-VCD) with pre-defined simulated videos. Therefore, the near-perfect results obtained on these datasets were somehow not convincing. As a result, in this paper, we introduce an effective real-scenario video copy detection system which aims to effectively and efficiently detect complex real video copies. Our system obtains decent results on a real-scenario large-scale video copy database (VCDB) generated in 2014, and measures the trade-off between effectiveness and efficiency. We believe our work can be regarded as the beginning for this challenging problem.