Caries or holes in the teeth, can be difficult to detect. Although X-rays have been illuminated since the late 1800s, there are factors that make the method vulnerable.

- Windows with a lot of light input and computer screens that are not optimal affect the X-rays, says Henrik Wennergren, IT manager and project manager at Folktandvården Stockholm.

That is why Folktandvården Stockholm is now testing a whole new approach to get even more secure results.

The public dental office tests neural networks

The new technology is so-called neural networks. Using an algorithm that is input, millions of images can be analyzed in a network. There, the algorithm tells the system what is right and wrong and what results you want. In this way, AI, artificial intelligence, should be able to locate and eliminate sources of error.

- It works much like the brain, you learn the network by entering more information. The system teaches what is caries and what is not caries, says Henrik Wennergren.

The new system will be introduced this fall

So far, the algorithm's hit accuracy is around 80 percent. In the autumn when the image analysis is complete and the desired result is achieved, the result should then be examined by X-ray specialists. Then you want to test the technology out at the clinics.

- All industries are looking at AI right now, so it will explode and there is as much potential as possible in all industries to use these technologies.