网络图像检索重排序系统建模和仿真

基于关键词的网络图像检索得到的结果往往是海量且无序的,给用户造成不便。本文构建了一个网络图像检索结果重排序系统并进行了仿真。提出了一种利用图像SIFT局部特征构建图像间视觉Graph结构对图像进行聚类的重排序算法,并引入检索关键词的网络翻译以提高检索效率。仿真结果证明,系统的重排序结果满足网络用户的检索期望,验证了算法的有效性。     

Semantic guided knowledge graph for large-scale zero-shot learning

Zero-shot learning has received growing attention, which aims to improve generalization to unseen concepts. The key challenge in zero-shot tasks is to precisely model the relationship between seen and unseen classes. Most existing zero-shot learning methods capture inter-class relationships via a shared embedding space, leading to inadequate use of relationships and poor performance. Recently, knowledge graph-based methods have emerged as a new trend of zero-shot learning. These methods use a knowledge graph to accurately model the inter-class relationships. However, the currently dominant method for zero-shot learning directly extracts the fixed connection from off-the-shelf WordNet, which will inherit the inherent noise in WordNet. In this paper, we propose a novel method that adopts class-level semantic information as a guidance to construct a new semantic guided knowledge graph (SG-KG), which can correct the errors in the existing knowledge graph and accurately model the inter-class relationships. Specifically, our method includes two main steps: noise suppression and semantic enhancement. Noise suppression is used to eliminate noise edges in the knowledge graph, and semantic enhancement is used to connect two classes with strong relations. To promote high efficient information propagation among classes, we develop a novel multi-granularity fusion network (MGFN) that integrates discriminative information from multiple GCN branches. Extensive experiments on the large-scale ImageNet-21K dataset and AWA2 dataset demonstrate that our method consistently surpasses existing methods and achieves a new state-of-the-art result.     

Dynamic Waveband Switching in WDM Mesh Networks Based on a Generic Auxiliary Graph Model

Waveband switching (WBS) is a promising technique to reduce the switching and transmission costs in the optical domain of a wavelength-division multiplexing (WDM) network. This paper considers the problem of provisioning dynamic traffic using WBS in a WDM mesh network. The network of interest is a homogeneous WBS network, where each node has the functionality of WBS. The problem is called the dynamic WBS problem, which involves searching waveband-routes or wavelength-routes for the dynamically arriving traffic requests. The objective is to minimize the total switching and transmission costs in the optical domain. To solve the dynamic WBS problem, an auxiliary graph model capturing the network state is proposed. Based on the auxiliary graph, two heuristic on-line WBS algorithms with different waveband grouping policies are proposed, namely the wavelength-first WBS algorithm based on the auxiliary graph (WFAUG) and the waveband-first WBS algorithm based on the auxiliary graph (BFAUG). Simulation results indicate that WBS is an attractive technique which reduces the overall switching and transmission costs by up to 30% in the network. The results also show that the WFAUG algorithm outperforms the BFAUG algorithm in terms of port savings and cost savings.     

Interference Mitigation and Joint Decoding for Uplink LDPC-Coded Relay Cooperation

This paper proposes an efficient inter- ference mitigation and joint decoding scheme for uplink LDPC-coded relay cooperation over a Rayleigh fading channel, where a concatenation of Minimum-mean-squared error linear detectors （MMSE） and BP-based joint itera- tire decoding based on the introduced treble-layer Tanner graph are effectively designed to filter and decode the cor- rupted received sequence at a base station. It is demon- strated by theoretical analysis and numerical simulations that the proposed design can well combine the gains from coding and diversity, which consequently leads to a sig- nificant performance improvement over the conventional cooperation system under the same conditions.     

Design of reliable fiber‐based distribution networks modeled by multi‐objective combinatorial optimization

In this work, we investigate the planning of hybrid fiber/very high bit rate digital subscriber line (HFV) access networks. This HFV variant is considered as an alternative to passive optical access network. It consists of implementing optical fibers in the access network part from the central office to the street cabinets, whereas the very high bit rate digital subscriber line technology is used in the last meter part, from street cabinets to subscribers. Firstly, we discuss the different tasks of this planning problem. We model this planning problem as a multi‐objective optimization problem, where different conflicting objectives have to be optimized at the same time. We consider network costs and network reliability. This problem is solved by two different approaches: the classical single‐objective optimization (SOO) and the multi‐objective optimization (MOO), which is a recent optimization approach that is gaining an increasing interest in practical optimization problems. MOO solves the problem by searching for different optimal trade‐offs between the optimization objectives. The performances of SOO and MOO are analyzed and compared with each other using three network instances. Copyright © 2012 John Wiley & Sons, Ltd.     

Homology analysis of malware based on graph

Malware detection and homology analysis has been the hotspot of malware analysis.API call graph of malware can represent the behavior of it.Because of the subgraph isomorphism algorithm has high complexity,the analysis of malware based on the graph structure with low efficiency.Therefore,this studies a homology analysis method of API graph of malware that use convolutional neural network.By selecting the key nodes,and construct neighborhood receptive field,the convolution neural network can handle graph structure data.Experimental results on 8 real-world malware family,shows that the accuracy rate of homology malware analysis achieves 93%,and the accuracy rate of the detection of malicious code to 96%.     

P-Zone： Connectivity-based Packet Path Tracing in Wireless Sensor Networks

In large-scale Wireless sensor networks （WSNs）, the network status is complex and unpredictable, which brings great challenges to practical network design and management. Tracing the route path of each data packet in the network is an important way to observe network behaviors and understand network dynamics. However, tracing the full route path of each packet could be highly challenging, due to the hard resource con- straint in WSNs. Our previous work proposes a hash-based path tracing mechanism, and leverages network connectivity and node locations to reduce the computational complexity. However, the node locations may be unavailable in some scenarios. In this work, we further propose a location-free enhancement to the hash-based path tracing mechanism, called P-Zone. P-Zone requires only network connectivity information to reduce the computational complexity. Theoretical analysis and practical simulations are conducted to evaluate the effectiveness and performance of our design. The results indicate that P-Zone can significantly reduce the computational complexity of the hashbased path tracing mechanism, while effectively tracing the full route path of each packet in the network in a real-time manner, and outperforms the state-of-the-art methods.     

Hiding color watermarks in halftone images using maximum-similarity binary patterns

This paper presents a halftoning-based watermarking method that enables the embedding of a color image into binary black-and-white images. To maintain the quality of halftone images, the method maps watermarks to halftone channels using homogeneous dot patterns. These patterns use a different binary texture arrangement to embed the watermark. To prevent a degradation of the host image, a maximization problem is solved to reduce the associated noise. The objective function of this maximization problem is the binary similarity measure between the original binary halftone and a set of randomly generated patterns. This optimization problem needs to be solved for each dot pattern, resulting in processing overhead and a long running time. To overcome this restriction, parallel computing techniques are used to decrease the processing time. More specifically, the method is tested using a CUDA-based parallel implementation, running on GPUs. The proposed technique produces results with high visual quality and acceptable processing time.     

基于顺序图着色方法的地面数字电视频率指配

介绍了频率指配的数学模型和顺序图着色方法,并举例说明了顺序图着色方法在地面电视频率指配规划中的实际应用.     

Optimal resource allocation scheme for cognitive radio networks with relay selection based on game theory

In this paper,we present a non-transferable utility coalition graph game(NTU-CGG) based resource allocation scheme with relay selection for a downlink orthogonal frequency division multiplexing(OFDMA) based cognitive radio networks to maximize both system throughput and system fairness.In this algorithm,with the assistance of others SUs,SUs with less available channels to improve their throughput and fairness by forming a directed tree graph according to spectrum availability and traffic demands of SUs.So this scheme can effectively exploit both space and frequency diversity of the system.Performance results show that,NTU-CGG significantly improves system fairness level while not reducing the throughput comparing with other existing algorithms.