Skill Memory in Biped Locomotion

Robots must be able to adapt their motor behavior to unexpected situations in order to safely move among humans. A necessary step is to be able to predict failures, which result in behavior abnormalities and may cause irrecoverable damage to the robot and its surroundings, i.e. humans. In this paper we build a predictive model of sensor traces that enables early failure detection by means of a skill memory. Specifically, we propose an architecture based on a biped locomotion solution with improved robustness due to sensory feedback, and extend the concept of Associative Skill Memories (ASM) to periodic movements by introducing several mechanisms into the training workflow, such as linear interpolation and regression into a Dynamical Motion Primitive (DMP) system such that representation becomes time invariant and easily parameterizable. The failure detection mechanism applies statistical tests to determine the optimal operating conditions. Both training and failure testing were conducted on a DARwIn-OP inside a simulation environment to assess and validate the failure detection system proposed. Results show that the system performance in terms of the compromise between sensitivity and specificity is similar with and without the proposed mechanism, while achieving a significant data size reduction due to the periodic approach taken.     

Relational Learning with GPUs: Accelerating Rule Coverage

Relational learning algorithms mine complex databases for interesting patterns. Usually, the search space of patterns grows very quickly with the increase in data size, making it impractical to solve important problems. In this work we present the design of a relational learning system, that takes advantage of graphics processing units (GPUs) to perform the most time consuming function of the learner, rule coverage. To evaluate performance, we use four applications: a widely used relational learning benchmark for predicting carcinogenesis in rodents, an application in chemo-informatics, an application in opinion mining, and an application in mining health record data. We compare results using a single and multiple CPUs in a multicore host and using the GPU version. Results show that the GPU version of the learner is up to eight times faster than the best CPU version.     

Visually impaired people and the emerging connected TV: a comparative study of TV and Web applications’ accessibility

With the TV signal digitization and the current market growth of connected TVs, the authors envision the appearance of accessibility barriers to visually impaired persons. The paper addresses the following hypothesis: (1) visually impaired users want to extend their TV usage to explore new TV features; (2) TV applications are in less conformance with accessibility guidelines compared to their Desktop versions. Additionally, the authors wanted to assess whether guideline conformance reflected real TV accessibility problems experienced by users. The methods used for this study included surveys aimed at characterizing the interest of the visually impaired population regarding the use of TV, and specifically of Web applications on TV, an automated accessibility evaluation to compare TV and Desktop versions of the same Web application, to understand their conformance with accessibility guidelines, and a user study where participants with visual impairments were asked to perform some tasks on both versions. From the survey, we confirmed that people with visual disabilities are interested in extra features on their TV. Results from the automated accessibility evaluation show that TV applications are in a significantly better level of conformance with accessibility guidelines. The user study has illustrated that users were unable to complete any task using the TV versions of the applications. The results from these studies demonstrated that the new features that come with connected TVs still have a long way to go in order to be accessible by all. Furthermore, they lead us to concur with other works that automated evaluations are not enough to assess the accessibility of a Web page.     

Active learning paradigms for CBIR systems based on optimum-path forest classification

This paper discusses methods for content-based image retrieval (CBIR) systems based on relevance feedback according to two active learning paradigms, named greedy and planned. In greedy methods, the system aims to return the most relevant images for a query at each iteration. In planned methods, the most informative images are returned during a few iterations and the most relevant ones are only presented afterward. In the past, we proposed a greedy approach based on optimum-path forest classification (OPF) and demonstrated its gain in effectiveness with respect to a planned method based on support-vector machines and another greedy approach based on multi-point query. In this work, we introduce a planned approach based on the OPF classifier and demonstrate its gain in effectiveness over all methods above using more image databases. In our tests, the most informative images are better obtained from images that are classified as relevant, which differs from the original definition. The results also indicate that both OPF-based methods require less user involvement (efficiency) to satisfy the user's expectation (effectiveness), and provide interactive response times.     

Satellite-based estimates of groundwater storage variations in large drainage basins with extensive floodplains

This study presents monthly estimates of groundwater anomalies in a large river basin dominated by extensive floodplains, the Negro River Basin, based on the synergistic analysis using multisatellite observations and hydrological models. For the period 2003-2004, changes in water stored in the aquifer is isolated from the total water storage measured by GRACE by removing contributions of both the surface reservoir, derived from satellite imagery and radar altimetry, and the root zone reservoir simulated by WGHM and LaD hydrological models. The groundwater anomalies show a realistic spatial pattern compared with the hydrogeological map of the basin, and similar temporal variations to local in situ groundwater observations and altimetry-derived level height measurements. Results highlight the potential of combining multiple satellite techniques with hydrological modeling to estimate the evolution of groundwater storage.     

IsoGeometric analysis using T-splines on two-patch geometries

We develop optimal approximation estimates for T-splines in the case of geometries obtained by gluing two standard tensor product patches. We derive results both for the T-spline space in the parametric domain and the mapped T-NURBS in the physical one. A set of numerical tests in complete accordance with the theoretical developments is also presented.     

Diagnosis of lung nodule using Moran’s index and Geary’s coefficient in computerized tomography images

This paper analyzes the application of Moran’s index and Geary’s coefficient to the characterization of lung nodules as malignant or benign in computerized tomography images. The characterization method is based on a process that verifies which combination of measures, from the proposed measures, has been best able to discriminate between the benign and malignant nodules using stepwise discriminant analysis. Then, a linear discriminant analysis procedure was performed using the selected features to evaluate the ability of these in predicting the classification for each nodule. In order to verify this application we also describe tests that were carried out using a sample of 36 nodules: 29 benign and 7 malignant. A leave-one-out procedure was used to provide a less biased estimate of the linear discriminator’s performance. The two analyzed functions and its combinations have provided above 90% of accuracy and a value area under receiver operation characteristic (ROC) curve above 0.85, that indicates a promising potential to be used as nodules signature measures. The preliminary results of this approach are very encouraging in characterizing nodules using the two functions presented.

虚拟空间会议系统中视频合成技术的研究与实现

虚拟空间会议系统能够把所有与会者合成同一个虚拟会场中,从而构造一个能提供无缝的会场环境,眼神接触、凝视感知等普通多媒体会议系统不具备的虚拟协同工作环境。VST系统是我们刚刚完成的一个虚拟空间会议系统。视频合成技术是其实现的关键技术之一。本文介绍了VST系统中视频合成技术的研究与实现。     

Algorithms for the Homogeneous Set Sandwich Problem

A homogeneous set is a non-trivial module of a graph, i.e. a non-empty, non-unitary, proper subset of a graph's vertices such that all its elements present exactly the same outer neighborhood. Given two graphs the Homogeneous Set Sandwich Problem (HSSP) asks whether there exists a sandwich graph which has a homogeneous set. In 2001 Tang et al. published an all-fast algorithm which was recently proven wrong, so that the HSSP's known upper bound would have been reset thereafter at the former determined by Cerioli et al. in 1998. We present, notwithstanding, new deterministic algorithms which have it established at We give as well two even faster randomized algorithms, whose simplicity might lend them didactic usefulness. We believe that, besides providing efficient easy-to-implement procedures to solve it, the study of these new approaches allows a fairly thorough understanding of the problem.