3D Semantic Scene Completion: A Survey

International Journal of Computer Vision - Semantic scene completion (SSC) aims to jointly estimate the complete geometry and semantics of a scene, assuming partial sparse input. In the last years...     

Differential evolution solution to the SLAM problem

A new solution to the Simultaneous Localization and Modelling problem is presented in this paper. The algorithm is based on the stochastic search for solutions in the state space to the global localization problem by means of a differential evolution algorithm. This non linear evolutive filter, called Evolutive Localization Filter (ELF), searches stochastically along the state space for the best robot pose estimate. The set of pose solutions (the population) focuses on the most likely areas according to the perception and up to date motion information. The population evolves using the log-likelihood of each candidate pose according to the observation and the motion errors derived from the comparison between observed and predicted data obtained from the probabilistic perception and motion model.The proposed SLAM algorithm operates in two steps: in the first step the ELF filter is used at local level to re-localize the robot based on the robot odometry, the laser scan at a given position and a local map where only a low number of the last scans have been integrated. In the second step, the aligned laser measures and the corrected robot poses are used to detect whether the robot is revisiting a previously crossed area (i.e., a cycle in the robot trajectory exists). Once a cycle is detected, the Evolutive Localization Filter is used again to estimate the accumulated residual drift in the detected loop and then to re-estimate the robot poses in order to integrate the sensor measures in the global map of the environment.The algorithm has been tested in different environments to demonstrate the effectiveness, robustness and computational efficiency of the proposed approach.     

A Faà di Bruno Hopf algebra for a group of Fliess operators with applications to feedback

A Faà di Bruno type Hopf algebra is developed for a group of integral operators known as Fliess operators, where operator composition is the group product. Such operators are normally written in terms of generating series over a noncommutative alphabet. Using a general series expansion for the antipode, an explicit formula for the generating series of the compositional inverse operator is derived. The result is applied to analytic nonlinear feedback systems to produce an explicit formula for the feedback product, that is, the generating series for the Fliess operator representation of the closed-loop system written in terms of the generating series of the Fliess operator component systems. This formula is employed to provide a proof that local convergence is preserved under feedback.     

Admission control for a responsive distributed middleware using decision trees to model run-time parameters

The software in modern systems has become too complex to make accurate predictions about their performance under different configurations. Real-time or even responsiveness requirements cannot be met because it is not possible to perform admission control for new or changing tasks if we cannot tell how their execution affects the other tasks already running. Previously, we proposed a resource-allocation middleware that manages the execution of tasks in a complex distributed system with real-time requirements. The middleware behavior can be modeled depending on the configuration of the tasks running, so that the performance of any given configuration can be calculated. This makes it possible to have admission control in such a system, but the model requires knowledge of run-time parameters. We propose the utilization of machine-learning algorithms to obtain the model parameters, and be able to predict the system performance under any configuration, so that we can provide a full admission control mechanism for complex software systems. In this paper, we present such an admission control mechanism, we measure its accuracy in estimating the parameters of the model, and we evaluate its performance to determine its suitability for a real-time or responsive system.     

Face tracking with automatic model construction

This paper describes an active model with a robust texture model built on-line. The model uses one camera and it is able to operate without active illumination. The texture model is defined by a series of clusters, which are built in a video sequence using previously encountered samples. This model is used to search for the corresponding element in the following frames. An on-line clustering method, named leaderP is described and evaluated on an application of face tracking. A 20-point shape model is used. This model is built offline, and a robust fitting function is used to restrict the position of the points. Our proposal is to serve as one of the stages in a driver monitoring system. To test it, a new set of sequences of drivers recorded outdoors and in a realistic simulator has been compiled. Experimental results for typical outdoor driving scenarios, with frequent head movement, turns and occlusions are presented. Our approach is tested and compared with the Simultaneous Modeling and Tracking (SMAT) [1], and the recently presented Stacked Trimmed Active Shape Model (STASM) [2], and shows better results than SMAT and similar fitting error levels to STASM, with much faster execution times and improved robustness.     

Stochastic automata-based estimators for adaptively compressing files with nonstationary distributions.

This correspondence shows that learning automata techniques, which have been useful in developing weak estimators, can be applied to data compression applications in which the data distributions are nonstationary. The adaptive coding scheme utilizes stochastic learning-based weak estimation techniques to adaptively update the probabilities of the source symbols, and this is done without resorting to either maximum likelihood, Bayesian, or sliding-window methods. The authors have incorporated the estimator in the adaptive Fano coding scheme and in an adaptive entropy-based scheme that "resembles" the well-known arithmetic coding. The empirical results obtained for both of these adaptive methods are obtained on real-life files that possess a fair degree of nonstationarity. From these results, it can be seen that the proposed schemes compress nearly 10% more than their respective adaptive methods that use maximum-likelihood estimator-based estimates.     

A comparison of directed formulations for the capacitated minimal spanning tree problem

In this paper, we compare linear integer programming directed formulations for the capacitated minimal spanning tree (CMST) problem. This problem is directly related to network design and consists of finding the cheapest way to link a set of terminals at different locations with a central node (a computer site, for instance). The present work focuses on a single commodity flow formulation presented in [3,4]. This formulation includes two sets of variables and, hence, it also must include coupling constraints between the two sets of variables. These coupling constraints can be defined in several ways. The main result of this work establishes that when the strongest form of the coupling constraints is used in the flow formulation (as in [4]), the equivalent formulation using only theX _ij variables satisfies the so-called multistar constraints. These constraints are shown to be stronger than the weakest form of the generalized subtour breaking constraints and, hence, our result improves on the characterization given in [5] for the flow formulation presented in [4]. For certain parameters, the multistar constraints induce facets of the non-directed CMST polytope (see [2]). We also show that it is not possible to compare the linear relaxations of a formulation that involves the multistar constraints with one that involves the generalized subtour breaking constraints.     

Second Life as a support element for learning electronic related subjects: A real case

Looking for more active and motivating methodological alternatives from the students’ perspective, which promote analysis and investigation abilities that make the student a more participative agent and some learning processes are facilitated, a practical study was conducted in the University of La Sabana (Chía, Colombia), in Computing Engineering (INF) and Agroindustrial Production Engineering (IPA) academic programs under the guiding question: Is Second Life an attractive tool for students, does it increment motivation, promote participation and facilitate learning processes in electronic related subjects in Engineering degrees?Second Life (SL) platform was then used as a traditional on-site class complement in electronic related subjects. A follow up was conducted with several groups during a year and a half, some students followed the subject with the traditional on-site methodology and others did it following a mixed methodology that combined traditional on-site lessons with virtual sessions trough SL, making a comparison and applying a quantitative as well as a qualitative methodology for the information analysis and recollection.After the specified study period it was found that using SL as part of the methodology, increments motivation toward electronic related subjects, and promotes participation and investigation, due to the fact that this tool makes learning possible in a funnier, more interactive and deeper way than if it had been taught in a master class. Having real-time access from a virtual classroom to all types of audiovisual information, helps more senses get involved in the learning process, making it more effective; nevertheless a weakness was detected related to the attention and concentration: students tend to get distracted, because their classroom is the computer and from there they have access not only to the many worlds of SL but to all the social networks and the Internet in general.     

A collaborative platform for cognitive decision making in the Knowledge Society

Two key aspects of the Knowledge Society are the interconnection between the actors involved in the decision making processes and the importance of the human factor, particularly the citizen’s continuous learning and education. This paper presents a new module devoted to knowledge extraction and diffusion that has been incorporated into a previously developed decision making tool concerning the Internet and related with the multicriteria selection of a discrete number of alternatives (PRIOR-Web). Quantitative and qualitative procedures using data and text mining methods have been employed in the extraction of knowledge. Graphical visualisation tools have been incorporated in the diffusion stage of the methodological approach suggested when dealing with decision making in the Knowledge Society. The resulting collaborative platform is being used as the methodological support for the cognitive democracy known as e-cognocracy.