Contemporary sport is becoming unbelievably digitized. It is being promoted and highlighted by social media and TV. Basically, all offline sport activities go into online. However, sport organizations, teams, coaches forget that the increased by the efforts of media interest to the sport, makes them correspond to enormous demands of audiences. Therefore, it is obvious that analytical solutions for teams and media promotion & engagement tools for channels or sport federations based on the screening of events by computer vision software with artificial intelligence may help to accomplish higher scores and outreaches of audiences.
Automatic generation of comments on volleyball
SpdLoad covers needs for sport organizations, teams and media companies in the area of:
- Business Analysis;
- UX/UI design;
- Development of AI systems for sport events visual screening, players performance analysis and comments generation.
Benefits of working with us:
- We have a solid knowledge for the development of integrative computer vision apps which use neural networks while making decisions. It was gained during the development of our own computer vision system;
- We provide services that help our clients to collect genuine stats and score data which help to measure activities’ KPIs and players’ performance;
- Our solution can be integrated into training systems of other teams, broadcasting software.
Our team has taken part in the development of sport events screening tool and comment generation neural network for one of our current clients.
The client is actually sport association from USA. They want to replace a position of stats coach with an automated tool for collecting game stats and generation of comments.
For them it was extremely hard to find a team which is experienced in integration of the computer vision technology into sport. Besides, hourly rates for AI developers in the USA starts from $120/hour.
The solutions is actually a software, application for the automation of stats collecting and analyzing during a volley match.
There are only 2 alternatives on the market. First one is a truly academic research which has no commercial application potential. The second one is adapted for other sports contrariwise to ours which is focused on volleyball.
The solution was realized using 3 different neural networks. Each of them tracks the coordinates of a particular object. Then algorithms compute coordinates and send them into the interface via API while the system is deployed in a cloud.