A step towards the use of writer's properties for text recognition

The reading activity, which is considered as an information processing system, involves the integration of different processes, knowledge, and modalities in order to transform a graphical representation into a symbolic representation. McClelland and Rumelhart (1981) have introduced the concept of interactivity in their human reading model, called the interactive activation model. In this model, the visual word recognition (perception) is realised by the interaction between three different abstraction levels (feature, letter, word). We show that this concept of interactivity can be extended by introducing an interaction within the grapheme level. This is possible only in the context of handwritten text recognition where the morphological invariants of a handwriting can be extracted. Due to these invariants (writer's properties), the interactions between and within different abstraction levels can be guaranteed and allow the automatic reading system to adapt itself to each handwriting.