Search

Most of the methods for real-time semantic segmentation do not take into account temporal information when working with video sequences. This is counter-intuitive in real-world scenarios where the main application of such methods is, precisely, being able to process frame sequences as quickly and accurately as possible. In this paper, we address this problem by exploiting the temporal information provided by previous frames of the video stream. Our method leverages a previous input frame as well as the previous output of the network to enhance the prediction accuracy of the current input frame. We develop a module that obtains feature maps rich in change information. Additionally, we incorporate the previous output of the network into all the decoder stages as a way of increasing the attention given to relevant features. Finally, to properly train and evaluate our methods, we introduce CityscapesVid, a dataset specifically designed to benchmark semantic video segmentation networks. Our proposed network, entitled FASSVid improves the mIoU accuracy performance over a standard non-sequential baseline model. Moreover, FASSVid obtains state-of-the-art inference speed and competitive mIoU results compared to other state-of-the-art lightweight networks, with significantly lower number of computations. Specifically, we obtain 71% of mIoU in our CityscapesVid dataset, running at 114.9 FPS on a single NVIDIA GTX 1080Ti and 31 FPS on the NVIDIA Jetson Nano embedded board with images of size 1024 × 2048 and 512 × 1024 , respectively. [ABSTRACT FROM AUTHOR]

Current video content providers require a huge demand for network resources to obtain the best quality of experience from the end-users perspective. This measure can be achieved if the network can anticipate itself to different existing issues (bandwidth problems, etc.). Due to traditional Internet Protocol networks cannot handle this problem due to the rigidity of its architecture, software defined networks is introduced as a potential answer for resolving the existing challenges in video content delivery decoupling the control plane and data plane. Utilizing the unified perspective of the network, it is possible to develop a monitoring tool obtaining the network metrics to be used afterwards. In this paper, we propose the application of different clustering algorithms for monitoring software defined networks to validate the number of statistics queries decrements using the data and error rates to improve the network traffic and to reduce the overload. The simulations demonstrate that the expectation-maximization algorithm highly reduces the monitoring queries maintaining the best data and error rate values in the video delivery. [ABSTRACT FROM AUTHOR]

A jurisprudence search system is a solution that makes available to its users a set of decisions made by public bodies on the recurring understanding as a way of understanding the law. In the similarity of legal decisions, jurisprudence seeks subsidies that provide stability, uniformity, and some predictability in the analysis of a case decided. This paper presents a proposed solution architecture for the jurisprudence search system of the Brazilian Administrative Council for Economic Defense (CADE), with a view to building and expanding the knowledge generated regarding the economic defense of competition to support the agency's final procedural business activities. We conducted a literature review and a survey to investigate the characteristics and functionalities of the jurisprudence search systems used by Brazilian public administration agencies. Our findings revealed that the prevailing technologies of Brazilian agencies in developing jurisdictional search systems are Java programming language and Apache Solr as the main indexing engine. Around 87% of the jurisprudence search systems use machine learning classification. On the other hand, the systems do not use too many artificial intelligence and morphological construction techniques. No agency participating in the survey claimed to use ontology to treat structured and unstructured data from different sources and formats. [ABSTRACT FROM AUTHOR]

This paper addresses the problem of designing trojan detectors in neural networks (NNs) using interactive simulations. Trojans in NNs are defined as triggers in inputs that cause misclassification of such inputs into a class (or classes) unintended by the design of a NN-based model. The goal of our work is to understand encodings of a variety of trojan types in fully connected layers of neural networks. Our approach is: (1) to simulate nine types of trojan embeddings into dot patterns; (2) to devise measurements of NN states; and (3) to design trojan detectors in NN-based classification models. The interactive simulations are built on top of TensorFlow Playground with in-memory storage of data and NN coefficients. The simulations provide analytical, visualization, and output operations performed on training datasets and NN architectures. The measurements of a NN include: (a) model inefficiency using modified Kullback–Liebler (KL) divergence from uniformly distributed states; and (b) model sensitivity to variables related to data and NNs. Using the KL divergence measurements at each NN layer and per each predicted class label, a trojan detector is devised to discriminate NN models with or without trojans. To document robustness of such a trojan detector with respect to NN architectures, dataset perturbations, and trojan types, several properties of the KL divergence measurement are presented. [ABSTRACT FROM AUTHOR]

The development of digital media, the increasing use of social networks, the easier access to modern technological devices, is perturbing thousands of people in their public and private lives. People love posting their personal news without consider the risks involved. Privacy has never been more important. Privacy enhancing technologies research have attracted considerable international attention after the recent news against users personal data protection in social media websites like Facebook. It has been demonstrated that even when using an anonymous communication system, it is possible to reveal user's identities through intersection attacks or traffic analysis attacks. Combining a traffic analysis attack with Analysis Social Networks (SNA) techniques, an adversary can be able to obtain important data from the whole network, topological network structure, subset of social data, revealing communities and its interactions. The aim of this work is to demonstrate how intersection attacks can disclose structural properties and significant details from an anonymous social network composed of a university community. [ABSTRACT FROM AUTHOR]

Increasingly sophisticated antivirus (AV) software and the growing amount and complexity of malware demand more processing power from personal computers, specifically from the central processor unit (CPU). This paper conducted performance tests with Clam AntiVirus (ClamAV) and improved its performance through parallel processing on multiple cores using the Open Multi-Processing (OpenMP) library. All the tests used the same dataset constituted of 1.33 GB of data distributed among 2766 files of different sizes. The new parallel version of ClamAV implemented in our work achieved an execution time around 62% lower than the original software version, reaching a speedup of 2.6 times faster. The main contribution of this work is to propose and implement a new version of the ClamAV antivirus using parallel processing with OpenMP, easily portable to a variety of hardware platforms and operating systems. [ABSTRACT FROM AUTHOR]

Social network analysis aims to obtain relational data from social systems to identify leaders, roles, and communities in order to model profiles or predict a specific behavior in users' network. Preserving anonymity in social networks is a subject of major concern. Anonymity can be compromised by disclosing senders' or receivers' identity, message content, or sender-receiver relationships. Under strongly incomplete information, a statistical disclosure attack is used to estimate the network and node characteristics such as centrality and clustering measures, degree distribution, and small-world-ness. A database of email networks in 29 university faculties is used to study the method. A research on the small-world-ness and Power law characteristics of these email networks is also developed, helping to understand the behavior of small email networks. [ABSTRACT FROM AUTHOR]

Mobile Ad Hoc Networks (MANETs) are multihop wireless networks of mobile nodes without any fixed or preexisting infrastructure. The topology of these networks can change randomly due to the unpredictable mobility of nodes and their propagation characteristics. In most networks, including MANETs, each node needs a unique identifier to communicate. This work presents a distributed protocol for dynamic node IP address assignment in MANETs. Nodes of a MANET synchronize from time to time to maintain a record of IP address assignments in the entire network and detect any IP address leaks. The proposed stateful autoconfiguration scheme uses the OLSR proactive routing protocol for synchronization and guarantees unique IP addresses under a variety of network conditions, including message losses and network partitioning. Simulation results show that the protocol incurs low latency and communication overhead for IP address assignment. [ABSTRACT FROM AUTHOR]

The TCP/IP protocol allows the different nodes in a network to communicate by associating a different IP address to each node. In wired or wireless networks with infrastructure, we have a server or node acting as such which correctly assigns IP addresses, but in mobile ad hoc networks there is no such centralized entity capable of carrying out this function. Therefore, a protocol is needed to perform the network configuration automatically and in a dynamic way, which will use all nodes in the network (or part thereof) as if they were servers that manage IP addresses. This article reviews the major proposed auto-configuration protocols for mobile ad hoc networks, with particular emphasis on one of the most recent: D2HCP. This work also includes a comparison of auto-configuration protocols for mobile ad hoc networks by specifying the most relevant metrics, such as a guarantee of uniqueness, overhead, latency, dependency on the routing protocol and uniformity. [ABSTRACT FROM AUTHOR]

The applications of wireless sensor networks comprise a wide variety of scenarios. In most of them, the network is composed of a significant number of nodes deployed in an extensive area in which not all nodes are directly connected. Then, the data exchange is supported by multihop communications. Routing protocols are in charge of discovering and maintaining the routes in the network. However, the appropriateness of a particular routing protocol mainly depends on the capabilities of the nodes and on the application requirements. This paper presents a review of the main routing protocols proposed for wireless sensor networks. Additionally, the paper includes the efforts carried out by Spanish universities on developing optimization techniques in the area of routing protocols for wireless sensor networks. [ABSTRACT FROM AUTHOR]

The article examines genetic variation of the FMR1 gene among four Mexican populations of Mestizo, Huichol, Purepecha and Tarahumara. Molecular analysis of the FMR1 gene showed 30 different alleles among the 525 chromosomes evaluated. The increased number of the unusual allele of 32 repeats observed in the Mexican mestizo population can be explained from its frequency in at least two Mexican native populations.