Make the world a better place while playing Grand Theft Auto.

You might think that playing Grand Theft Auto only gives you the chance to steal random cars or run over people for no reason, but the truth is you could be making the world a better place to be. How’s that possible, you might ask?

Well, scientists from Darmstadt University in Germany and Intel Labs have developed an intelligent method to import data from GTA to computers to teach self-driving vehicles operate better on the road.

Researchers say data from modern computer games can be “almost as good, or sometimes even better” than real data, and using it is much cheaper than gathering real-world visual data. Developers have created special software, capable of classifying different objects on roads within the game.

They have already extracted 25,000 frames from the computer game with varying weather conditions at different times of day. Then, they annotated them in 49 hours in total – a process that would have taken days with rear-world data. The process is explained in the video above.

The gathered data is used for teaching autonomous cars the difference between pedestrians and other objects when navigating through the real world, TheNextWeb says.

The better an autonomous car is, the safer it is on the road – there’s no question about that. But that’s not all – the smarter a self-driving car is, the smarter it behaves in traffic, reducing the time we spend in traffic jams. And, generally speaking, the more self-driving cars in traffic, the more efficient the traffic gets.

Meanwhile, while waiting for the researchers to teach autonomous cars and make them better, you can spend some time in GTA teaching yourself how to make tricks – just like this guy makes stunts, involving a semi truck and well-placed C4. Good luck!

Source: Visual Inference via TheNextWeb

Photo: YouTube screenshot