Jean-Yves Hervé

Building Detection and Recognition for an Automated Tour Guide

We present a building recognition module for an automated tour guide system which, instead of using user input or artificial landmarks, relies solely on real-time video recorded from a wearable camcorder to recognize faces of buildings and localize the user. Strong vertical edges are first detected on each frame to suggest the existence of building objects. Additional information such as color feature is collected to eliminate edges from trees, cars, and pedestrians. Vertical edges, in addition to horizontal edges, are further processed to locate window frames and thereafter faces of a building. An image section identified as a building face is then rectified for perspective distortion and recognized by a neural network. We train the neural network using whole and partial building face images before the recognition takes place.

Adaptive Target Detection and Matching for a Pedestrian Tracking System

We present a 3D tracking system for detecting and tracking multiple targets. An extended Kalman Filter (EKF) is used to maintain each targets 3D state and provide location predictions to the pattern matchers whose task it is to follow the targets in images. An adaptive background modeling algorithm is used together with the tracking process to detect moving objects in complex environments. We propose a warping-based pattern matching approach to deal with object deformation during tracking. We present examples of results of our tracker for outdoor scenes.

A Pattern for the Integration of Conceptual Models in Support of Multidisciplinary Efforts to Develop Software

Conceptual models are an essential part of research, especially where multidisciplinary teams need multiple models to capture and convey the disciplinary views. Here the authors present a reusable pattern that describes an effective means of operating within a multidisciplinary group to develop software. Patterns for repeatable processes provide savings in time. The quality of resulting products is higher, for the patterns have previously been used and the “bugs” have been “ironed out”. Examples here are from an active project in which the authors are developing simulation software for social scientists and security experts to analyze building evacuations. The team is comprised of engineers, computer scientists and social scientists. The paper gives justification for the use of patterns and a description of this process pattern in the context of the evacuation project. The process pattern includes: project description, construct identification, diagramming, construct mapping, spiral model for process iteration, interdisciplinary literacy, and special issues in the software environment.

System for game enhancement using scene recognition in emulated consoles

This paper investigates the modification of an existing game console emulator utilizing scene recognition to perform audio-visual enhancement of vintage video games. Such games featured low-resolution graphics with limited colors and low-quality electronic music. Information contained about the virtual hardware state information is exploited to provide high-resolution, full-color graphics and a recorded soundtrack, complete with atmospheric effects, in real-time using the original game logic. The system developed as part of this research uses a modified version of the FCEUX emulator for the Nintendo Entertainment System. The system is capable of loading a commercially-released game, marking scenes for later recognition, mapping scenes to audio-visual events, and recognizing those scenes to enable those events. The result is an immersive game experience that far exceeds what was possible in the original game hardware.

Casual gaming as a means to raise awareness of vector borne disease risks

Vector-borne diseases cause hundreds of thousands of people around the world to suffer debilitating illnesses each year; illnesses such as Lyme Disease, Babesiosis, tick borne encephalitis, Crimean-Congo Hemorrhagic Fever, and Rocky Mountain spotted fever. The spread of these diseases is caused by human-biting ticks. However, in virtually every case there is a preventive strategy to stop ticks from attaching to humans and therefore preventing infection. Lack of knowledge of these prevention strategies, and awareness of ticks-diseases infection risks contribute to a steady increase in infection rates. In the recent years, there has been a growing popularity in casual gaming and interactive media in society, especially among women aged 35+ and younger children around the age of 13. Coincidentally, this demographic also represents a significant portion of the population at most at risk of tickborne disease infection. We intend to capitalize on this demographic coincidence, creating casual games and applications that will reach this audience to develop better tickborne disease awareness and make prevention fun, easy, and popular. In this paper we present several educational casual games and multimedia applications we have developed using the game engine Unity3D, Flash, and other web technologies. We also discuss our overall strategy of using New Media as a tool to educate the general public by offering educational content on different platforms and devices, from the Web to the iPhone.

Using a Game Engine to Integrate Experimental, Field, and Simulation Data for Science Education: You Are the Scientist!

The purpose of this project is to use a game engine to integrate geo-referenced research data, whether experimental or simulated, to present it in an active, interactive form to the user. The data being geo-referenced means that every image, video, or sound file, every pressure map, and every simulated temperature chart is attached to a specific point on a map or body. These data may even be time-referenced, so that different data sets may be available at the same location for different times of the day or seasons of the year. Target users for such interactive applications are high-school and college students who can then conduct their own “experiments” or “explorations” as a way to get exposed to the problems and methodologies of science and research. We use two examples of projects to illustrate the approach.

Agent Abilities in a Landmark-based Mapping Model

We present a landmark-based model for agents to uniquely learn a partial map of an unknown environment. Our definition of landmark recognition and agent abilities allow different agents to construct different maps. We define landmarks in terms of feature sets that tie the various agent abilities into the landmark recognition process. The ultimate goal of having agents learn partial and unique maps is to promote different behavioral outcomes from higher level planning. Our implementation focuses on providing unexpected and believable Non-Player Characters (NPC) in modern games.

Adaptive, Region-based, Layered Background Model for Target Tracking

Background subtraction is a method frequently used in target detection and tracking applications. In this paper, we present a new approach to background subtraction that relies on a two-layer background model. This approach uses a single intensity model for all the pixels rather than separate models for individual pixel, by exploiting the spatial dependency of pixels. It models the background without modeling the foreground, and treats background updating and tracking as tightly integrated dynamic processes rather than consecutive static modules. This method can handle illumination changes, permanent changes to the background, and targets that stop moving and later start moving again. Example results are presented