Top Data Science detects human poses from video with their algorithm

How could we use city CCTV cameras more efficiently in situations that are potentially threatening? Smart Urban Security and Event Resilience Project (SURE) and Top Data Science carried out an agile experiment and found some promising answers in Tampere, Finland.

The name – Top Data Science – summarises the main idea of the company. It offers co-creative problem solving with the state-of-the-art AI and machine learning, usually for long-term customers. The company has been developing, for example, AI solutions for video analytics.

– A SURE agile experiment was an opportunity to use existing technology for safety and security needs and in a real urban environment, says Timo Heikkinen, Top Data Science CEO.

Agile experiment proved it works

Top Data Science chose a promising algorithm capable of analysing human poses from a video feed. The City of Tampere has a network of surveillance cameras in the city centre. One of them was used in the experiment.

In addition, SURE partnership was required: the algorithm was deployed on the Nokia Scene Analytics platform ja results were send to Insta Blue Aware platform.

– Our algorithm is able to detect human poses and activities, even if there is a lot of people in the frame. It visualises the humans with stick figures and also counts them, says Heikkinen.

It’s always been the SURE policy to respect information security and privacy. Therefore AI will be used to detect humans from video feed, but never to recognise individual people.

For better urban safety and security

The SURE project is aiming to integrate data, gathered from various sensors, into situational awareness – a real-time view on what is happening in the centre of Tampere.

For Top Data Science, the next step might be choosing and analysing a designated activity. Running, falling, lying on the ground or violence are all activities that the algorithm is capable of detecting. This would lead to improvements in urban security as it could mean quicker emergency responses.

Careful groundwork, smoothly running experiment

The SURE project had already done their groundwork thoroughly before they launched the agile experiments. Petri Nykänen, Director at Business Tampere, explains that preparations involved choosing target beneficiaries and recognising their security challenges; also examining technical system requirements with the partners.

From Top Data Science’s point of view, the SURE agile experiment was certainly worthwhile. It was a showcase of what the algorithm can do in a real urban environment – and a unique chance to get feedback.

– There were numerous security and emergency services professionals in attendance, when we presented our technology. The experiment will also be a very good reference for us in the global and growing field of urban security, says Heikkinen.

The experiment was a part of the SURE project (Smart Urban Security and Event Resilience). The project was granted 3,2 million euros from the Urban Innovative Actions, an ERDF-funded initiative of the European Commission.

Business Tampere is responsible for the upscaling of solutions, and provides a framework in which the technology providers can run the experiments.

Tampere Region has a significant concentration of safety and security know-how and organizations, both nationally and internationally. There is more than 250 organisations related to safety and security in the region. Tampere region safety and security cluster, which has been in operation since 2011, provides an easy access to local ecosystem.