Copenhagen-based startup Veo created an AI-powered camera for professional football (soccer for American readers) coaches and players to better understand and accordingly further develop their games and strategies.
The Denmark football camera can be placed midway on the touchline of a football field, and it will be able to cover the entire field. It is equipped with the company’s object-detection technology for tracking the ball in order to more insightfully catch the gameplay.
The Nvidia Veo mission is to provide professional video analytics capabilities globally, as said by Founder and CEO Henrik Teisbaek. The startup has already eyes on over 100,000 amateur clubs around the world. Teisbaek wants to pull the coaches off the chalkboards and get digitized to become professionals.
The device made by the Danish startup is triangular and bears a lime-green body, with branding on top. There are two 4K video cameras to sit on a tripod. Captured data gets transferred to Veo’s cloud service.
The panoramic videos from the collected footages processed by the cloud system will provide enough space for coaches and players to analyze. Coaches can even placemarks on the field to analyze in the videos right after.
Veo started operating in 2016 and is presently working with 20 football clubs of Copenhagen. The first generation model of the camera can capture about 100 matches per month. Teisbaek pointed out that it’s four to five times of what’s broadcasted by Danish television!
The second-gen model is in works and will launch later this year, as Veo says. The Veo World cup camera is priced at about $1,200 Euros. The company reports that around 100 clubs have already bought the second-gen model in pre-sales.
For video processing, Veo has its convolutional neural network working, which has a training of 1.5 million images. This huge bundle features about 1 million footballs and 500,000 players. Fast-find game highlights for scored goals and more is one capability the device already possesses, with Veo focusing on its further growth.
Teisbaek said that they use NVIDIA processors for accelerating the processing of roughly 500 million pixels per second. He added that these are requirements for real-time processed videos. The company’s other usage includes GPUs from a German data center and Nvidia’s cuDNN with TensorFlow.
Surely, the elite class of the football world from Samara Arena is getting Veo’s updates today! Denmark was leading 1 goal to nil against Australia in their second group match in 2018 FIFA World Cup while writing this report!