81–85 of 85 posts
Some time ago I needed ball video tracking for a demo at work. I failed to find any out of the box solution which worked satisfactorily. It was way easier to find solutions for seemingly more complex tasks of people tracking or human pose estimation, for which you have numerous publically available algorithms/models.
Balls dont have many unique attributes, and move a lot faster. Sonetimes humans are easier:)
Earlier quoted context omitted.
Is that an auto-generated comment? I can't make sense of it at all.
Assume the marketing for the system is true ( https://www.pixellot.tv/ ) It's a modular technology. Just a camera and 4/5G card. It's all then just cloud software. They already have 10,000 installations. And I don't seem why any school would not have one. Sport is worth billions. The money is there to keep this rolling. 90,000 current hours per month are produced. This is a TV station run by software.
Earlier quoted context omitted.
Assume the marketing for the system is true ( https://www.pixellot.tv/ ) It's a modular technology. Just a camera and 4/5G card. It's all then just cloud software. They already have 10,000 installations. And I don't seem why any school would not have one. Sport is worth billions. The money is there to keep this rolling. 90,000 current hours per month are produced. This is a TV station run by software.
I think it's a neat approach for low-budget sports broadcasting but not very novel. Computer vision master students have been implementing ball tracking algorithms in first year courses since years. You usually combine some kind of feature detector (in this case probably a convolutional neural network, thus the "AI" naming) with some kind of state tracker like a Kalman filter. Getting the kinks out of the system is t…
Why do you think this? This is a funny example but they have millions of hours behind them.
> Computer vision master students have been implementing ball tracking algorithms in first year courses since years.
Well, I'm sure they do in a lab perhaps. Ball tracking algorithms and software will be 50 years old.
Pixellot has commercialised it and employed 150 people on this real world problem. This is the hard bit (the 1% of the 95%, 5%, 1%) Not the Facebook someone programmed over the weekend.
Either train more with bald players, or do better at choosing the 'thing that is being converged upon'. Humans would be able to follow a game regardless of the shape or color of the ball's appearance.
I imagine for soccer the task of determining where the ball is based on player locations would be a lot harder than you think.
Earlier quoted context omitted.
Then have a few “known good” people and then average their actions with the actions of everyone else doing it – so long as you do “virtual panning” on a cropped image to give the controllers sufficient feedback, that'll give you enough data to kick out trolls / misbehavers in a “Web of Trust” type manner, and even out most other issues that occur.
While I'm sure that would work, that sounds way more complicated than using computer vision to track the ball.