I. Introduction
Great number of documents in several libraries contains a diversity of languages with mix variable writing types (machine-printed and handwritten). Despite the important improvement made over the last decade, the implementation of a text recognition system intended to recognize any text image extracted from a multi-script and multi-lingual document is always difficult. Hence, automatic writing type and script identification solutions are crucial to identify the handwritten and the machine-printed types and the used scripts in a document before feeding them to the suitable text recognition system.