How to build a tiny intelligent camera that can be placed anywhere and see for us? Eyes of Things (EoT) is developing a platform for embedded computer vision that will allow innovative products and services.

Eyes of Things

This is a guest blog post written by Oscar Deniz, EoT project coordinator

Computer vision going massive

Computer vision is a discipline where we use computer software to analyse images for content. Traditionally focused on factory automation (inspect parts, measure distances, …all in carefully controlled factory conditions) , this field is now booming with applications everywhere: videogames, automotive, drones, wearable headsets, automatic surveillance, intelligent toys and companions, etc.

In EoT we want to develop a flexible open platform to contribute to this trend. Our tiny intelligent camera is targeted at original equipment manufacturers (OEMs), the idea being to provide the right tool to develop visually intelligent products and services with short time-to-market.

First steps…

The project is now into its 8th month and a lot of work has been done at multiple levels. Training into the advanced technologies began even before the official starting date of the project, with ‘bootcamps’ held by project partners. On a technical level, the successive iterations of the physical device are now underway (see some prototypes in the project Flickr account). Essentially, the idea is to squeeze the fundamental components (camera, processor, WiFi) into a pen-drive sized component. Hardware development runs in parallel to the development of associated software, which shall run both in the device and in external controlling computers (smartphones, tablets or PCs). To get a better idea, have a look at our video!

Smile, you’re being watched…

Any device that captures personal data is a sensitive issue. Legislation is clear, for example, in the fact that the picture of a person is considered personal data. Far from being an obstacle, this has made the project a lot more exciting and open to multidisciplinary collaborations. Ethics experts were recruited to follow the project’s evolutions and will have an active role in ensuring that platform trials do not trespass any privacy barrier. Privacy is also being taken into account on a technical level, and a number of technical decisions have already been made to ensure that privacy can be preserved.

Come and see us at ICT 2015 and other events

Overall, EoT is an exciting experience in which we build a powerful device from the ground up. The project is attracting a lot of interest from both academia and industry, with partners eager to see the final device. Of course, not everything is technical work. A big effort is being made in dissemination and networking. The project coordinators are also well-aware of the social role of this EU-funded effort, and aspects like awareness, public engagement and gender issues are being considered throughout.

The first ‘public demo’ of EoT is expected for September 2015 at the 9th International Conference on Distributed Smart Cameras (Seville, Spain). This will be followed by ICT 2015 - Innovate, Connect, Transform - where we expect to show ‘real devices doing real things’. There, we will also try to convey the potential of computer vision. We are also focusing on increasing the network of collaborators through EU-funded calls. Take a look at our entry in the Exhibition Catalogue and don’t forget to visit our stand located in the Transform Area.

Site under construction

Work continues in the project with exciting developments coming up.. In the next months we will continue to release new versions that are closer to the final device both in physical aspect and software capabilities. The most exciting time will probably come during the second half of the project, in which we want to demonstrate the platform in 4 application scenarios. However, that’s another story, for another blog!

Get involved!

EoT has been funded by you, so we’d like to have your feedback. There are in fact many ways in which you can contribute to the objectives of EoT. Just by learning a little about computer vision (what is it, what it is achieving…) you will be already contributing to our goals. We have also set up a way in the project website to receive your encouraging comments, criticisms or desired features. Follow #EyesOfThings developments in Twitter and of course keep an eye (for lack of a better term) on what we do in the project website.