Abstrakt
Future multi-media technologies are expected to support on-line processing of huge amounts of high-dimensional data without any special pre-processing. Growing Neural Networks designed for efficient image processing also involve data-dependent adjustment of both the number and position of the neurons that improves generalization.
In this way, local features detected automatically by Growing Neural Networks impact a transparent and compact representation of the extracted knowledge. The performed case studies refer to face and hand-written digit recognition.