The IMRA experiment transformed the way viewers experience a marathon event!

The IMRA application aimed to enrich the viewer’s marathon experience, by providing personalized content on demand. Viewers in such events, are usually interested in one or more athletes. However, it is impossible to track their effort throughout the race. IMRA made this possible!

The IMRA solution comprises of two parts: the IMRA app and the IMRA backend service. The IMRA application is available at the Google Play at this link.By leveraging the crowdsourcing trend of the viewers, a smart summarization video including the athlete of their interest is available on demand. IMRA utilizes the state-of-the-art Deep Learning methods for video processing, person detection and athlete’s bib number identification in order to provide the ability to watch your favourite athlete’s effort throughout the race, without having to move at all.

IMRA has been deployed in Bristol infrastructure and it had a great acceptance from the participants of the Marathon race.

Using the IMRA app
Using the IMRA app

The main outcomes of the trial are the following:

  • Created engaging content for marathon spectators. Building a smart-video summarization of the race, personalized according to the user’s request.
  • Achieved the best possible QoE for the end-users. Deliver video content efficiently with the minimum interactions and lowest possible initial playout time.
  • Experimented with different configurations. D-Cube experimented with different deployment plans, placing the individual components to the edge or to the data center, in order to conclude to the most optimal Service Function Chain.
  • Compared streaming performance between FLAME and Cloud solution. Deploying on a cloud platform and comparing the QoE against deploying on a FLAME platform.
  • Optimized Deep-Learning-based Video processing algorithms for the edge nodes of a 5G adaptive platform. Leveraging the FLAME platform’s features and the state-of-the-art Deep Learning optimization libraries in order to reduce processing time at the edge.
  • Developed a working prototype that could lead to the IMRA launch in the market. Demonstrate the capabilities of a tested prototype to potential clients and acquire feedback directly from the Market.
  • Provided feedback to the FLAME consortium. Benchmarked FLAME platform and provided suggestions regarding its enhancement.
  • Familiarize with 5G platforms. Acquire knowledge and experience with a innovative 5G platform.