Computer vision for mobile computing and interaction

The field of mobile applications represents one of the main potentials in information technology. Processing power and quality of imaging capabilities of mobile devices has been steadily increasing. Research in mobile computer vision has thus gained prominence, and there is an ever increasing number of research groups engaged in the problem of mobile vision. The work on the project has affirmed our group with the most pervasive groups in this area.

We were able to develop computer vision methods suitable for mobile devices, and use them to implement designed interaction scenarios in prototype applications.The use of computer vision makes for a very intuitive interaction with mobile device, greatly simplifying it. Prototype applications in addition to visual information use various other modes of sensory data, e.g. location (GPS) to shrink the space of possible object matches, and orientation data for controlling the interval of processing and thus the energy consumption. There are few comprehensive and complete implemented systems in existence, despite overall progress in the development of algorithms, and increased processing power. The rare ones are largely limited to interaction with well-defined image data, e.g. to identify text, faces, bar codes, or advertising boards. Our system promises stable operation even in case of free use in urban environments.

Enhancing HIP recognition with robust geometric identification and hierarchical processing represents an original scientific contribution, while the implemented demonstrators act as a solid basis for evaluating the possible ways of interaction.