Unsupervised Learning of Human Action Categories Using Spatial-Temporal Words

We present a novel unsupervised learning method for human action categories. A video sequence is represented as a collection of spatial-temporal words by extracting space-time interest points. The algorithm automatically learns the probability distributions of the spatial-temporal words and the intermediate topics corresponding to human action categories. This is achieved by using latent topic models such as the probabilistic Latent Semantic Analysis (pLSA) model and Latent Dirichlet Allocation (LDA). Our approach can handle noisy feature points arisen from dynamic background and moving cameras due to the application of the probabilistic models. Given a novel video sequence, the algorithm can categorize and localize the human action(s) contained in the video. We test our algorithm on three challenging datasets: the KTH human motion dataset, the Weizmann human action dataset, and a recent dataset of figure skating actions. Our results reflect the promise of such a simple approach. In addition, our algorithm can recognize and localize multiple actions in long and complex video sequences containing multiple motions.     

Motion Planning for Haptic Guidance

Haptic devices allow a user to feel either reaction forces from virtual interactions or reaction forces reflected from a remote site during a bilateral teleoperation task. Also, guiding forces can be exerted to train the user in the performance of a virtual task or to assist him/her to safely teleoperate a robot. The generation of guiding forces relies on the existence of a motion plan that provides the direction to be followed to reach the goal from any free configuration of the configuration space (-space). This paper proposes a method to obtain such a plan that interleaves a sampling-based exploration of -space with an efficient computation of harmonic functions. A deterministic sampling sequence (with a bias based on harmonic function values) is used to obtain a hierarchical cell decomposition model of -space. A harmonic function is iteratively computed over the partially known model using a novel approach. The harmonic function is the navigation function used as motion plan. The approach has been implemented in a planner (called the Kautham planner) that, given an initial and a goal configuration, provides: (a) a channel of cells connecting the cell that contains the initial configuration with the cell that contains the goal configuration; (b) two harmonic functions over the whole -space, one that guides motions towards the channel and another that guides motions within the channel towards the goal; and (c) a path computed over a roadmap built with the free samples of the channel. The harmonic functions and the solution path are then used to generate the guiding forces for the haptic device. The planning approach is illustrated with examples on 2D and 3D workspaces. This work was partially supported by the CICYT projects DPI2005-00112 and DPI2007-63665.     

Line patterning of graphite and the fabrication of cheap, inexpensive, “throw-away” sensors

A new simple method (“line patterning technique”) using only standard office equipment is described whereby clearly defined, electrically conducting patterns of graphite can be deposited on polymer (plastic) or paper substrates. The properties of the conductive patterns have been characterized by electrical conductivity and SEM measurements. Sensors were constructed by using interdigitated patterns of graphite deposited on plastic and paper, and coated with a thin film of conducting electronic polymer, e.g. polyaniline emeraldine salt.     

‘Podracing’: experimenting with mobile TV content consumption and delivery methods

Recently, mobile TV has been launched in several countries. While mobile TV integrates television contents into mobile phones, the most personal of communication devices, it becomes interesting to know how this feature will be used throughout the day and in varying contexts of everyday life. This paper presents empirical results on the use of mobile TV with different delivery mechanisms and both quantitative and qualitative results on how end-users prefer to use mobile TV contents in different situations. The data is based on ongoing empirical research in Finland in 2006 and 2007. The mobile TV services under study included both news and entertainment contents, and were tested in 3G, DVB-H and Wi-Fi networks using different delivery paradigms: broadcast, on-demand and download. To explore the use of different delivery methods and content consumption, we have developed a mobile TV service protoype, called Podracing. The analysis shows that users appreciated up-to-date information and information-rich media forms and contents especially for mobile news delivery. There was high demand for only the latest news on mobiles. The real-time property was considered important. Most of the users looked at the headlines or followed the news several times a day – much more often than the traditional TV and news prime times would allow.     

Visualization of Simulated Urban Spaces: Inferring Parameterized Generation of Streets, Parcels, and Aerial Imagery

Urban simulation models and their visualization are used to help regional planning agencies evaluate alternative transportation investments, land use regulations, and environmental protection policies. Typical urban simulations provide spatially distributed data about number of inhabitants, land prices, traffic, and other variables. In this article, we build on a synergy of urban simulation, urban visualization, and computer graphics to automatically infer an urban layout for any time step of the simulation sequence. In addition to standard visualization tools, our method gathers data of the original street network, parcels, and aerial imagery and uses the available simulation results to infer changes to the original urban layout and produce a new and plausible layout for the simulation results. In contrast with previous work, our approach automatically updates the layout based on changes in the simulation data and thus can scale to a large simulation over many years. The method in this article offers a substantial step forward in building integrated visualization and behavioral simulation systems for use in community visioning, planning, and policy analysis. We demonstrate our method on several real cases using a 200 GB database for a 16,300 km2 area surrounding Seattle, Washington.     

Markovian-based traffic modeling for mobile ad hoc networks

Mobile ad hoc networks (MANETs) show very significant difference with respect to other computer networks due to the presence of extremely large packet loss bursts. The development of protocols for mobile ad hoc networks, especially multimedia protocols, require extensive evaluation either through simulation or real-life tests. Such testing consumes a great amount of resources both in terms of time and trace file sizes. Therefore, finding efficient means of reducing the amount of data that is stored and processed is quite important to accelerate the evaluation of different audio/video streaming applications. If, moreover, we are able to model the loss pattern experienced, we can further accelerate the evaluation process.In this work we propose two models based on hidden Markov chains that are able to grasp both packet arrivals and packet loss patterns in MANETs. A simpler two-state model is proposed to model losses when proactive routing protocols are used, while a more complex three-state model is proposed for reactive routing protocols. We also introduce a new set for packet loss pattern measurements that can be of interest for the evaluation of audio/video streaming applications.Experimental results show that the proposed models can adequately reproduce extremely long packet loss patterns, typical of MANET environments, with a high degree of accuracy. Overall, we find that the proposed models are able to significantly reduce both the simulation time and the trace file sizes required.     

Soft computing and cooperative strategies for optimization

In the last years, metaheuristics have emerged as powerful algorithmic approaches which have been applied with great success to difficult combinatorial optimization problems. However, this does not mean that metaheuristics can be applied blindly to any new problem.In this contribution we showed how the most basic ingredients of Soft Computing, namely fuzzy sets and fuzzy rules, are used in the context of a simple metaheuristic and a cooperative strategy based on it, to obtain successful results for the p-median problem.     

A new control scheme for visual servoing

This research presents a new control scheme for visual servoing that takes into account the delay introduced by image acquisition and image processing. Firstly, previous control schemes for visual servoing are discussed and then a new control scheme is proposed and fully described. Afterwards, the capabilities (steady-state errors, stability margins, step time response, etc.) of the proposed control scheme and of previous ones are analytically analyzed and compared. Next, several simulations and experimental results are provided to validate the analytical results and to illustrate the benefits of the proposed control scheme. In particular, it is shown that this new control scheme clearly improves the performance of the previous ones.     

Acclimatization model of an aerobic bioreactor for the treatment of toxic wastewater

This work proposes a mathematical model for the acclimatization process of a bioreactor treating toxic wastewater. Experimental data was used to identify the changing kinetic parameters of the model as acclimatization progresses. It was found that only one key parameter, the specific biomass growth rate function, changed during the acclimatization process. Therefore, an acclimatization model was proposed to explain the changes of this parameter.     

Proteomic analysis in cerebrospinal fluid of patients with atypical nonketotic hyperglycinemia and pulmonary hypertension - A pilot study

A variant phenotype of nonketotic hyperglycinemia has been described by our group associated with pulmonary hypertension. The aim of this study is to investigate the cerebrospinal fluid proteomes to get an insight into this neurodegenerative process producing leukoencephalopathy with white matter spongiform degeneration. DIGE and MALDI-TOF-TOF analyses were performed to carry out the proteomic study of four patients against three normal controls and one additional control of a classical nonketotic hyperglycinemia. The differential proteomic analysis showed a displacement of some series of spots toward the acidic side. The shifted proteins showed a high degree of carbonylation and increased methionine sulfoxidation was found in cystatin C and in vitamin-D-binding protein. These findings in addition to the increase of serum malondialdehyde concentration provide evidence of an oxidative stress in the patients under study, which is probably systemic rather than mainly confined to the CNS. The similarities of our findings with those found in other neurodegenerative diseases suggest that oxidative damage is commonly involved in these pathologies. DIGE technology improves the 2-D PAGE differential analysis and it is suitable in proteomic studies with a small number of cases.