المجلة الدولية للعلوم والتقنية

The recognition of the online Arabic handwriting is a complex problem. Arabic is cursive and unconstrained language, in addition, the language uses diacritics with distinct shapes places above or underneath the characters. Several of these diacritics has an irregular geometric shapes such as hamza (ء). The online recognition typically traces the pen tip positions and its vertical movements pen-down or pen-up. This paper presents a new method of measuring the pen up and down direction to detect the diacritic and which character belongs to. We applied this method to identify the Arabic online handwriting. The results of the extensive experiments achieved a high accuracy rate for writer-dependent and writer-independent in detection the position and segmentation of the diacritics.