De grootste kennisbank van het HBO

Inspiratie op jouw vakgebied

Vrij toegankelijk

Terug naar zoekresultatenDeel deze publicatie

Greetings Door Opener

Rechten: Alle rechten voorbehouden

Greetings Door Opener

Rechten: Alle rechten voorbehouden

Samenvatting

The goal of the 'Greetings Door Opener' project is to create a proof of concept program that automatically greets a person entering the Fourtress office. This should be done using a camera and facial recognition software to give personalized greetings. The creation of facial detection and recognition software is not a part of the project and will be acquired externally.

The project was divided in three main phases, the first phase consisted mainly of researching a suitable face detection software package. This resulted in choosing OpenCV, an open source computer vision library, as the most suitable for the project. As a computer vision library various detection and recognition techniques could be used. For this project Haar-like feature detection and Eigen face recognition were selected.

The second phase was aimed at development and resulted in a standalone client that could detect and recognize faces, as well as some administrative features, like enrolling and deleting users.

The final phase was reserved for implementing extensions and enhancements. The main enhancement was to rewrite the program as a client/server setup. This in order to reduce the workload on the client pc's, as well as increase the flexibility of the overall program.

The project is successful as a proof of concept. Due to the modularity of the program the recognition and detection parts of the project can easily be updated or replaced. With the Client/Server approach a new stage for the project has been reached. This gives other developers easy access to the detection software without having to interface with a software library.

The next step would be to improve the recognition software, as the current recognition solution is rather sensitive to changes in facial expressions and head orientation of the detected person. Replacing it with either a commercial product or implementing a different technique in the current solution are two possible solutions.

Secondly, the detection speed with the current detection method is quite time consuming. At less than a frame per second the amount of suitable images for recognition is limited. Another improvement might be a higher resolution camera, or a different lens to increase the amount of detail to enhance the recognitions. This would however further slow down the detection speed.

Toon meer
OrganisatieAvans Hogeschool
AfdelingAIM Academie voor ICT en Media
AfstudeerorganisatieFourtress
Jaar2008
TypeBachelor
TaalEngels

Op de HBO Kennisbank vind je publicaties van 26 hogescholen

De grootste kennisbank van het HBO

Inspiratie op jouw vakgebied

Vrij toegankelijk