Skip to main content

Deep vision learning for swimmer safety and analytics

Periodic Reporting for period 1 - LYNXIGHT (Deep vision learning for swimmer safety and analytics)

Reporting period: 2019-02-01 to 2019-07-31

Lynxight's mission is to eradicate the problem of drowning, using advanced computer vision and deep learning technologies. Drowning is the second-leading cause of injury death for children under 14, responsible for 6,000 annual fatalities and 650 severe non-fatal injuries daily – all in the EU. The LYNXIGHT project is a swimmer safety management and analytics platform, deployed initially in pools and subsequently in other aquatic venues. We detect early signs of water distress, unusual swimmer behavior and a variety of pool analytics to enable proactive life-guarding and optimize operational resources. Our main objective is to develop a swimmer safety system, deployed anywhere humans enter water, to prevent all types of drowning and water-distress. Equally important, our aim is to make this system economically viable for any customer type, so it becomes widespread and effective for all facility types, sizes and affluence-levels.
Technically, our feasibility study now spans over 45 video feeds from a wide variety of swimming pool types, shapes and sizes, streaming data through the cloud to our R&D team for analysis and machine learning networks. We’ve extended our study to include a wider variety of pool types, and are maintaining a high level of performance on this comprehensive dataset, with our proprietary core computer vision capabilities continually delivering high quality and robust results. We’ve finalized camera requirements, and specifications for our video processing workstation, have completed the system and cloud architectures and finished developing the lifeguard smartwatch. For IPR, we’ve completed an extensive survey of patent literature, and freedom to operate is clear. Finally, GDPR and ISO20380 compliance processes are nearing completion.

Commercially, we’ve finished verifying a comprehensive product-market fit including main customer pain points and matching product features, means of sales, distribution and maintenance, business and pricing models and go-to-market strategy. In the EU, we’ve participated in several key industry events and met with leading pool industry companies and have a good understanding of the methods to reach and sell to different types of customers, their willingness to pay and payment models. We’ve secured several strategic partnerships with pool equipment and service providers and cooperation agreements with insurance agencies which will help us advance into the commercialization stage.
After the progress made in this phase one project as described in the above paragraph, we are now ready to move on to the final stage of completing our commercialization efforts. Technically, this means moving to TRL8/9 – completing algorithmic and software design and addressing several key issues which the current design is lacking in (e.g. performance in high swimmer density), and validating the system on a large number of pool types and scenarios. Commercially, this means completing our deployment and support plans and finalizing strategic partnerships with sales, distribution, and insurance partners. This effort will require another two years and 2-3M Euro, and we are will start to raise the required capital in the upcoming months. The end result will be an inexpensive system which can prevent the vast majority of drownings in pools, towards expansion to additional aquatic venues.
sample images from selected pilots
lifeguard smartwatch application
RTL Aktuell evening news 25.07.2019: