Object Recognition for the Internet of Things

Fonte Standard Gras
Identifiant

Unité de recherche

INNOSUISSE

Numéro de projet

9198.1;5 PFES-ES

Titre du projet

Object Recognition for the Internet of Things

Titre du projet anglais

Object Recognition for the Internet of Things

Textes relatifs à ce projet

	Allemand	Français	Italien	Anglais
Description succincte	-	-	-
Résumé des résultats (Abstract)	-	-	-

Textes saisis

Catégorie	Texte
Description succincte (Anglais)	Object Recognition for the Internet of Things
Résumé des résultats (Abstract) (Anglais)	We propose a CTI project between the partners ETH Zurich and kooaba. The mission of kooaba is to 'hyperlink' the real world with the digital world using the integrated camera of mobile phones. Pictures taken by users are recognised by our image recognition server and related digital content is returned to the phone. The goal of this project is to extend the abilities of the existing Software in scalability and breath of application. Extending scalabil ity requires research in high-dimensional indexing of visual features and integration of multimodal cues such as geographic position of objects and users. Extending the breadth of application is achieved by the integration of recognition for special types of objects such as text and barcodes in addition to general object recognition. The results of this research will lead to the release of a new product generation at kooaba.

Catégorie

Texte

Description succincte
(Anglais)

Object Recognition for the Internet of Things

Résumé des résultats (Abstract)
(Anglais)

We propose a CTI project between the partners ETH Zurich and kooaba. The mission of kooaba is to 'hyperlink' the real world with the digital world using the integrated camera of mobile phones. Pictures taken by users are recognised by our image recognition server and related digital content is returned to the phone. The goal of this project is to extend the abilities of the existing Software in scalability and breath of application. Extending scalabil ity requires research in high-dimensional indexing of visual features and integration of multimodal cues such as geographic position of objects and users. Extending the breadth of application is achieved by the integration of recognition for special types of objects such as text and barcodes in addition to general object recognition. The results of this research will lead to the release of a new product generation at kooaba.