MobiLearn

The MobiLearn project concerns the use of deep learning approaches to exploit images taken by mobile cameras, such as front cameras in motor vehicles.

The high level technical objectives of the MobiLearn project are as follows:

• Create, design and develop high performance analytical algorithms for analyzing mobile camera images.

• Create, design and develop robust deep learning methods for image processing, applicable for critical applications.

• Create demonstrators and prototypes capable of addressing industrial scenarios based on real and concrete application cases.

The project is organised around a number of application cases which are used, in the event of convincing research results, to develop new products and services in the form of innovative functions exploiting the images taken by the vehicle's front camera (and, where applicable, other sensors) to be integrated into navigation technologies for the driver's service (Advanced Driver Assistance Systems). In this project, the two partners Aisin and ACIC combine their specific expertise:

• Expertise in electronic and computer systems embedded in vehicles as well as in data analysis (big data and deep learning) for Aisin,

• Expertise in image analysis for ACIC.