Paint Multi-touch table
Authors
Tao, Tran ; Perona, Noémie ; Utasi, Tamás
Term
4. term
Publication year
2009
Abstract
This project builds on Aalborg University’s experimental multi-touch tabletop to address limitations in finger detection speed and accuracy. It improves hardware components—most notably infrared LED illumination—sets up the camera and projector, and performs camera–projector calibration. An image-processing pipeline implemented with OpenCV combines background subtraction, brightness enhancement, smoothing, binarization, hole removal, and size filtering to reliably track fingers. An entertaining drawing application (“Paint”) serves as a demonstrator of the tabletop’s capabilities. The report covers design, implementation, and usability testing, and proposes further improvements (e.g., enhanced lighting and two-camera setups). The excerpt states the goal of higher speed and accuracy but does not provide quantitative results.
Dette projekt tager udgangspunkt i Aalborg Universitets eksperimentelle multi-touch bord og adresserer mangler i hastighed og nøjagtighed ved fingerdetektering. Arbejdet omfatter forbedring af fysiske komponenter, især infrarød belysning med LED, opsætning af kamera og projektor samt kamera–projektor-kalibrering. Der implementeres en billedbehandlingspipeline i OpenCV med baggrundssubtraktion, lysstyrkeforbedring, udglatning, binarisering, hulfjernelse og størrelsesfiltrering for robust sporing af fingre. En underholdende tegneapplikation ("Paint") fungerer som demonstrator for systemets egenskaber. Rapporten beskriver design, implementering og brugbarhedstest samt foreslåede forbedringer (bl.a. bedre lys og brug af to kameraer). Uddraget angiver en målsætning om øget hastighed og nøjagtighed, men indeholder ikke kvantitative resultater.
[This apstract has been generated with the help of AI directly from the project full text]