Capsule networks are part of the world of artificial intelligence and are used particularly in the field of big data and smart data as well as in digital transformation. But what is behind them?
Put simply, capsule networks are a modern type of neural network that enables computers to recognise and understand images, texts or other data even better. In contrast to traditional methods, capsule networks not only analyse individual features, but also the relationships between these features. This makes them more robust against errors, for example if an object in an image is slightly distorted or partially obscured.
An illustrative example: conventional artificial intelligence can recognise a face in an image, but finds it difficult to distinguish the exact orientation or individual facial expressions. Capsule networks, on the other hand, not only identify that it is a face, but also recognise whether the person is smiling and where they are looking.
This opens up great potential for applications such as automatic image analysis, voice assistants or medical diagnoses. Capsule networks are therefore helping to make artificial intelligence even more precise and versatile.
Deutsch 
English (UK)