Method for configuring an image evaluation device and also image evaluation method and image evaluation device

An image evaluation device for determining an object type and an object sub-type of an imaged object the device comprising:

• one or more processors;
• and
• one or more memories coupled to the one or more processors for storing data to be processed by the one or more processors;
• wherein:
• the one or more processors are configured to:
• execute a first neural network module for recognizing image features in an image provided as input to the first neural network,
• execute a second neural network module for recognizing object types based on image features recognized by the first neural network module,
• select one of a plurality of sub-type recognition modules based on an object type recognized by the second neural network module, and
• execute a selected sub-type recognition module for recognizing object sub-types based on image features recognized by the first neural network module.

A non-transitory computer-readable storage medium having a computer program stored thereon that when executed by a processor causes the processor to:

• recognize image features in an image using a first neural network module of the computer program,
• derive an object type using a second neural network module of the computer program based on the image features recognized by the first neural network module,
• select a sub-type recognition module of the computer program specific to the object type derived by the second neural network module,
• determine an object sub-type using a selected sub-type recognition module based on the image features recognized by the first neural network module, and
• output the derived object type and the determined object sub-type.