Search results

Head detection in stereo data for people counting and segmentation

In this paper we propose a head detection method using range data from a stereo camera. The method is based on a technique that has been introduced in the domain of voxel data. For application in stereo cameras, the technique is extended (1) to be applicable to stereo data, and (2) to be robust with regard to noise and variation in environmental settings. The method consists of foreground selection, head detection, and blob separation, and, to improve results in case of misdetections, incorporates a means for people tracking. It is tested in experiments with actual stereo data, gathered from three distinct real-life scenarios. Experimental results show that the proposed method performs well in terms of both precision and recall. In addition, the method was shown to perform well in highly crowded situations. From our results, we may conclude that the proposed method provides a strong basis for head detection in applications that utilise stereo cameras.

Continuous Gait Velocity Analysis Using Ambient Sensors in a Smart Home

We present a method for measuring gait velocity using data from an existing ambient sensor network. Gait velocity is an important predictor of fall risk and functional health. In contrast to other approaches that use specific sensors or sensor configurations our method imposes no constraints on the elderly. We studied different probabilistic models for the description of the sensor patterns. Experiments are carried out on 15 months of data and include repeated assessments from an occupational therapist. We showed that the measured gait velocities correlate with these assessments.

Integrating meta-Information into recurrent neural network language models

Cameratoezicht voor valdetectie

Valongevallen zijn de meest voorkomende oorzaak van letsel door een ongeval bij ouderen. Omdat veel ouderen alleen wonen is het van belang dat de val tijdig wordt opgemerkt met een alarmsysteem. Er zijn veel systemen op de markt waarbij de gebruiker een alarmknop op een polsband of hanger moet indrukken. Uit onderzoek blijkt dat ouderen het alarmsysteem vaak niet gebruiken omdat de hanger niet meegedragen wordt of omdat de werking van het systeem niet duidelijk is. Een ‘ambient’ systeem heeft daarom de voorkeur. In het lectoraat Digital Life van de Hogeschool van Amsterdam bestuderen we ambient sensoren (schakelaars, stroommeters, drukmatten, bewegingsmelders etc.) en imaging sensors (camera’s, range camera’s, stereocamera’s) om activiteiten van mensen te meten. In onze presentatie beschrijven we het gebruik van ‘slimme’ camera’s voor valdetectie en laten enkele resultaten zien. Ook beschrijven we ons ‘ValLab’, waar onderzoekers en studenten HBO en WO data kunnen opnemen in realistische omgevingen.

Twice random, once mixed: applying mixed models to simultaneously analyze random effects of language and participants

Psychologists, psycholinguists, and other researchers using language stimuli have been struggling for more than 30 years with the problem of how to analyze experimental data that contain two crossed random effects (items and participants). The classical analysis of variance does not apply; alternatives have been proposed but have failed to catch on, and a statistically unsatisfactory procedure of using two approximations (known as F 1 and F 2) has become the standard. A simple and elegant solution using mixed model analysis has been available for 15 years, and recent improvements in statistical software have made mixed models analysis widely available. The aim of this article is to increase the use of mixed models by giving a concise practical introduction and by giving clear directions for undertaking the analysis in the most popular statistical packages. The article also introduces the djmixed add-on package for SPSS, which makes entering the models and reporting their results as straightforward as possible.

Increasing hosting capacity of low-voltage distribution network using smart charging based on local and dynamic capacity limits

While the Municipality of Amsterdam wants to expand the electric vehicle public charging infrastructure to reach carbon-neutral objectives, the Distribution System Operator cannot allow new charging stations where low-voltage transformers are reaching their maximum capacity. To solve this situation, a smart charging project called Flexpower is being tested in some districts. Charging power is limited during peak times to avoid grid congestion and, therefore, enable the expansion of charging infrastructure while deferring grid investments. This work simulates the implementation of the Flexpower strategy with high penetration of electric vehicles, considering dynamic and local power limits, to assess the impact on both the satisfaction of electric vehicle users and the business model of the Charging Point Operator. A stochastic approach, based on Gaussian Mixture Models, has been used to model different profiles of electric vehicle users using data from the Amsterdam public electric vehicle charging infrastructure. Several key performance indicators have been defined to assess the impact of such charging limitations on the different stakeholders. The results show that, while Amsterdam’s existing public charging infrastructure can host just twice the current electric vehicle demand, the application of Flexpower will enable the growth in charging stations without requiring grid upgrades. Even with 7 times more charging sessions, Flexpower could provide a power peak reduction of 57% while supplying 98% of the total energy required by electric vehicle users.